Informal meeting of the plenary to hear a briefing on the Independent Scientific Panel on AI - General Assembly, 80th session

Good morning.
Please take your seats.
I call to order the informal meeting of the plenary to hear briefing on the Independent International Scientific Panel on Artificial Intelligence.
In its paragraph one of Resolution 79 Stroke 325, the General Assembly established within the United Nations the multidisciplinary Independent Scientific Panel on Artificial Intelligence, including to provide updates on its work up to twice a year through an interactive dialogue of the plenary of the Assembly.
I warmly welcome you all to this meeting.
I also welcome very warmly the co chairs of the International Scientific Panel, miss Maria Ressa and Mr.
Yoshua Bengio.
We will start with the statement of the President of the General Assembly.
I will give a statement of the President of the General Assembly.
Excellencies, distinguished delegates.
It is my pleasure to open this interactive dialogue between the General Assembly and the co chairs of the Independent International Scientific Panel on Artificial Intelligence.
I'm delighted that so many are joining this meeting, underlining how important this topic is and also valuing the very important work of the panel and especially of its two co chairs.
Let me begin by extending my appreciation to all of you, but also to all different departments of the UN, underlining that this is really not only across regional but across sectoral issue also for the United Nations.
The appointment of this panel in February of this year reflected the growing recognition among member states of a simple truth.
To govern this powerful and rapidly evolving technology wisely, the international community needs expertise it can trust.
Expertise that is credible, independent, multidisciplinary, and grounded in evidence.
Even as we meet artificial intelligence is transforming almost every dimension of our lives, our economies, our societies, our public services, our classrooms, our hospitals at a rapid speed, providing on one hand, great opportunities to accelerate progress across the sustainable development goals to lift productivity and to make better decisions.
On the other hand, we all know at unprecedented speed how this technology can be turned against us.
Without proper guardrails, the same technology can pose profound risks to human rights, privacy, security, and dignity.
It can accelerate the spread of misinformation, widen digital divides, and enable new forms of exploitation, including gender based violence and sexual abuse through deep fakes.
It can also amplify hate speech.
As one young person warned during the commemorative meeting on the genocide in Rwanda the other day in the General Assembly Hall, hate speech now moved through screens, pixels, likes, and social media platforms, carrying, quote, old lies to new generations and threatening to turn their platforms and apps into battlefields of a young Rwandan teenager.
That warning should stay with us because the same technologies that they can connect and empower people can also spread hatred, intimidation, and division.
We've also seen growing concerns about the ways AI chat boards can affect young people, including reports of teenagers being encouraged towards suicide.
Indeed, one of the recurring lessons of human innovation is that technological progress has often moved faster than our collective ability to govern it well.
And this is exactly why this panel matters so much.
Member States need access to objective scientific assessment, the kind that separates evidence from speculation, that identifies opportunities and risks as they emerge, and that gives our policy discussions a foundation worth standing on.
The panel can be a bridge between the scientific community and the multilateral system, helping ensure that the choices we make here are informed by the best knowledge available to us.
It can do something more.
By offering impartial comprehensive analysis, the panel can also help narrow the divide between technological advanced nations and the global South, so that participation in a governance is not the privilege of a selected few, but the right shared by all.
And its work drawn on perspectives and expertise from every region of the world, must reflect the diversity if it is to serve us all.
Which brings me to why we are gathered in this particular chamber.
The General Assembly is the most universal and representative forum we have.
The one place where every member state, regardless of its size, its level of development, or its technological capacity, can help shape a common understanding of what our future and what our digital future and AI means to be.
No single country, no single company should write those rules alone.
Today's briefing is a step on that road.
It is an opportunity for you to engage directly with the two co chairs to learn firsthand about the panel's progress and just as importantly, for the panel to hear your views.
Under the mandate set out in Resolution 79 stroke 325, the co chairs will return to update this assembly up to twice a year.
This is therefore the beginning of a continuing conversation, not a one time exchange.
As we begin, I encourage you to make it an open and constructive one.
Ask the difficult questions, the challenges, the doubts, share your concerns, but also your hopes.
I would also like to encourage the co chairs of the scientific panel to be as candid as possible in their reflection also with the interaction with us, the United Nations and the member states.
As we all know how bureaucratic and slow this place sometimes can be and how fast the AI technology moves.
In this regards, I'm very much looking for this interaction today.
We are here to be honest and timely and evidence based together not only today, but also in the future.
With that, it's my great honor to give the floor to the two co chairs and I look forward to your briefing and the dialogue that will follow.
Thank you.
First, as we agreed, I now give the floor to Mr.
Joshua Bengio to provide, then it's wrong in my script.
So first, I give the floor to the lady.
I give the floor to miss Maria Ressa to revive the first part of the briefing.
Thank you.
Thank you.
Coach, thank you.
Distinguished member states, permanent observers and UN agencies, thank you.
These are two words I don't say lightly because when you created this panel, what you created matters more than I think any of us, maybe including you with the exception of President Berbach fully realized when you voted for it, you gave the world its first standing, independent scientific body on AI.
It's not a summit, it's not a communique.
It's a body with a mandate, a timeline, and accountability to evidence rather than to politics or profit.
You built the institution that the moment demands.
Thank you.
This is what we set out to build independent, science based, global in scope.
This panel exists to give the world something it does not yet have a shared, authoritative account of what AI means for humanity across seven critical domains with representation from every region of the world.
What we produced is a preliminary report, foundational by design, written in a compressed timeframe with, much more to follow.
We organized it not around academic convention, but around the questions the world is actually asking, and we built it to work at two levels at once.
For you, the member states, and the global dialogue who are our primary audience, but also for the public who are never far from our minds.
It's rigorous, but accessible.
And it's also honest about the limits of what we know and where the gaps remain.
Let me say something that's a little surprising.
The hardest part was not the differences between nations.
We found it remarkably easy to reach consensus across those lines.
On the overwhelming majority of issues, scientists from every region of the world, and that includes colleagues from across every divide that you can imagine, looked at the evidence and arrived at the same place.
Science gave us a common language that politics often can't.
This is why Tony Oke, he's here, who negotiated this panel into existence on behalf of the developing world, a block of 134 nations named our central question.
Can science stay independent of politics and of the industry that builds this technology? That's not an abstract question, not for him, not for us.
During the negotiations, independence proved contentious because the global South understands better than anyone what it means to depend on systems you did not build, cannot inspect, and do not control.
That gravity is real.
The pool of governments and the pull of a heavily industry funded field does not disappear because 40 experts sit around the table.
We felt it We have to be vigilant about it and we're working with it.
We have held the line, and we will keep holding it because the commitment of this panel is not to any company, not even to the countries we're from.
It is to each other and to the humanity we serve.
That's what independence means here.
It's the reason you can trust the report we will present in Geneva in July.
Now, consensus has a cost.
When for you know this better than anyone.
When 40 scientists from 40 different contexts must agree, you build toward the center.
The most contested findings require the most evidence and the most negotiation.
That is scientifically honest.
But please understand when you receive this report in July, what you are receiving is the floor of our concern, not the ceiling.
The evidence we agreed on is alarming enough and you heard some of it from President Baerbach.
What was this process really like? 40 people from 33 countries across five continents, physicists, philosophers, computer scientists, journalist, doctors, lawyers, economists, from the Philippines and Burkina Faso, from Russia and Chile, from Uganda and South Korea and Germany and Brazil, 40 people who didn't know each other really, who had never worked together appointed in February, expected to produce a scientifically rigorous assessment of the most consequential technology in human history by July.
Right? You know how much that takes.
Let me be honest.
There were moments of hair tearing frustration.
You ask 40 brilliant, opinionated, deeply committed people to reach consensus and you discover very quickly that brilliance does not make agreement easy.
It makes disagreement more precise.
But we did it.
The fact that we did 40 diverse minds across time zones, across disciplines, across radically different lived experiences is itself the message.
If science can hold together across these divides, then governance can too.
If we can build common ground, So can you.
This is what the panel proves, and I'm humbled by every single one of my colleagues, including and especially Joshua Benj, who you will hear from after me.
I've told him many times I love him.
It's been part of the process.
He's brilliant, he's as brilliant as he is principle as principled as he's brilliant.
The window to act is open.
This is what the report says.
That is what we are asking you to do.
40 of us.
Somehow, this process you put us in, we started as strangers and we are coming out as friends.
Joshua, please take them inside how our panel did this.
Thank you.
Thank you very much, Maria Ressa, indeed, Joshua Benj Pa for the second part of the briefing.
Thank you, Your Excellency, the President of the General Assembly, miss Anelea Berbach for convening this interactive dialogue.
Dear Excellencies and distinguished representatives, I really appreciate the opportunity to speak to you today to expand a bit on the beautiful speech that my co chair Mia Ressa just gave you on the work of our panel.
This moment in our collective history is being defined by an incredibly rapid technological transformation, the pace and scale of which are unprecedented.
I'm going to insist on this because from both a scientific and a societal standpoint, AI represents a turning point, a historical and transformative development because it is about the growing intelligence of machines and intelligence gives power.
With lots of risks that come with that, including a risk of excessive power concentration.
This is all happening faster than the speed at which societies appear able to adapt.
For those reasons, the importance of this panel's work and the urgency to act now given AI's current trajectory and the potential scenarios of where it will go in coming years, they cannot be overstated.
I firmly believe that science is a collaborative endeavor.
That's what I have lived throughout my life, my career.
An endeavor rooted in the search of truth.
That is the foundation.
That means objectivity and evidence regardless of the pressures and preferences of the world around it, which do matter, but it is your job.
Sometimes we have to say things that are not comfortable because the evidence points to it.
I see that same thinking reflected in the very title of this panel, the Independent International Scientific Panel on AI, independent of any company, government, or institution, even the UN itself.
International, as Maria said, given the ambitious scale of this global collaboration and the importance of having a global discussion because AI is going to be a global phenomenon, it already is.
Then of course, scientific to establish a rigorous shared base of evidence for policymakers.
The decisions made today regarding AI by governments around the world will thus have lasting consequences, and we may not have a lot of time to take the important decisions.
Our duty is to ensure those decisions are best informed using the highest standards of scientific integrity, which brings us to the process of pulling together this preliminary report to be released on July 1st.
We have had this year a very short time frame between the panel's appointment in February and the first expected output in Geneva in early July.
It was clear from the beginning that this first report could not be a comprehensive and exhaustive one.
In fact, I'm extremely proud and surprised by how much we were able to do and how much content you will find in the report.
To give this output the appropriate scientific rigor it deserves, we decided to deliver a preliminary assessment of the state of AI and its opportunities, impacts, and risks for society as set out in the resolution creating this panel.
You will find in this report four main sections.
First, defines AI, highlights how it differs from other emerging technologies, and addresses the urgent need for an independent evidence based assessment of AI systems.
The second part outlines the current scientific evidence highlighting different key assessments of topics including rapidly advancing capabilities, concentration of power, the AI divide, human and children rights, and many others.
The third section recaps key domain specific findings and use cases to illustrate these points.
The fourth focuses on evidence gaps, scope of the mandate, and the next steps for the panel.
The panel members all contributed to the developing of the documents structure, which was deemed best suited to fulfill our mandate within this deadline.
And we are really proud of the result.
It's scientific thoroughness is a result of a diversity of panelists backgrounds and contributions, which we were able to put to best use by forming seven complementary thematic working groups, which I will present.
These working groups were selected and identified by the panel as a whole and formed based on the most pressing policy relevant themes to be addressed by member states collaboration.
I'm going to highlight this question of policy relevant.
There are many things that are written in the scientific literature.
But we wanted to highlight the elements that matter for policy decisions to come.
The seven working groups themes are the following.
The first one on AI science advances, the science of AI itself.
It's advances, what the data shows, the trajectories which can help policymakers anticipate where it might go in coming years.
The second is about societal applications in other sciences, in health, for example, education, and agriculture.
The third is on economic implications.
Implications to labor, productivity, and development.
The fourth is on security, systems, and environmental implications.
The fifth is on human rights, information, and democracy.
The sixth is on cultural and individual flourishing, autonomy and child safety, and the last is on management practice and frameworks, capacity building and reliability.
In terms of process, the panelists got the chance to rank and choose the working group they felt best place to contribute their expertise to.
To avoid silos or overlap between working groups, we unanimously agreed to provide panelists with full access to other working groups drafts, meetings so everyone could go to all the meetings of the other working groups and decision making process.
Then within each working group, a facilitator was appointed on a volunteer basis again with the responsibility of keeping the groups on schedule, ensuring drafts were circulated and raising issues during our weekly plenary meetings.
In fact, often, someone else from the working group was presenting when that was most appropriate.
This facilitator role did not give an additional say to those people in the discussions, but helped reduce the administrative burden and make each group more productive and coordinated with the other groups.
Another panel subgroup was created based on organic discussions during our crucial in person meeting in Madrid in April.
It comprises of at least one volunteer from each working group, and it serves as an agile group that handles time sensitive editorial decisions, improves cross group communication, and helps ensure that no task falls to the wayside and goes unaddressed.
This agile group met weekly with meetings always open for any panel member to join if they so wished.
That sets the frequency of meetings for most panelists to two or three per week.
The working group meeting, the panel wide plenary, and especially near the end, a meeting of the agile editorial group for those involved.
I must say that I am incredibly grateful for the dedication and commitment of each panelist to contribute their time and expertise to these regular meetings.
The creation of these agile group as well as the crucial discussions on report structure and scope is a product of our incredibly valuable in person retreat that we held in Madrid, as I said, last April.
With less than three months ahead of us at that time, before the report deadline, this meeting came at a moment where time was of the essence and we needed to align quickly and efficiently.
This was an unequivocal success.
The productivity and collaborative spirit we created in person set the tone for the drafting of the preliminary report and it allowed an incredibly valuable flow of ideas, discussions, and collective problem solving.
Given these outcomes, it's clear that organizing further in person retreats will be vital for our future efforts.
It's necessary to continue funding such initiatives, please, and we appreciate any contribution you can make in this regard.
So we look forward to continuing our collaborative work to produce the panel's first annual report in 2027.
Until then, we will continue to provide a source for trustworthy evidence based assessments of AI as it continues evolving, notably with ad hoc theme briefs throughout the coming year.
We also look forward to refining, improving, and expanding our various processes, including by potentially setting up a mechanism for consultations with experts and stakeholders beyond the immediate panel.
I'm confident that this panel, which unites a diverse group of 40 of the world's foremost experts on AI will be a crucial voice in the global conversation on AI.
I look forward to continuing this work with my esteemed colleagues and I thank all of them for their exceptional contributions.
Thank you.
Thank you.
I thank Mr.
Benji.
I thank Maria Ressa for their briefings, not only informing about the further process, but underlining again that the decisions ahead are too important and too far reaching for any of us to make in isolation or to delay.
By this, I now open the floor for comments or questions.
This meeting will not have a pre established list of speakers.
Delegations wishing to take the floor are invited to press the microphone button.
Members are requested to limit their interventions to 5 minutes when speaking on behalf of a group and 3 minutes when speaking in a national capacity.
Time limit will be strictly enforced through an automatic microphone cutoff.
A timer will be projected on screen.
Those speaking on behalf of a group should approach the Secretariat to be given priority in the speaking order.
I would like to underline again that this is the unique opportunity, especially to bring forward your wishes and comments, but also use the opportunity to ask question.
As the first speaker, I now give the floor to Agai on behalf of the Group of 77 and China, followed by St.
Kitts and Neves and Saudi Arabia.
Thank you very much, Madam President and if I may, I will also speak in my national capacity for just a half a second to congratulate the panel for the presentation.
It's truly a pleasure to see the good vibes that we are seeing from the audience in your statement and in miss Maria's statement as well.
We are truly happy to see you there and to have such an impartial panel that can provide fair information at a moment when the world needs it.
Now I will read the G 70 seven's official statement.
Delegate, I'm honored to deliver this statement on behalf of the T 77 and China.
We have thanked the co chairs and Independent international Scientific Panel on Artificial Intelligence for this update on their work so far and we appreciate this opportunity to exchange views and interact with you in your capacity as representative of the 40 panels member to all of whom we extend our congratulations on your appointment.
The G 77 and China engaged actively in the negotiation of the Digital compact, as well as during the negotiation of Resolution 70 9325 on the term of reference and modalities for the establishment and functioning of the Independent International Scientific Panel on Artificial Intelligence.
The group has supported the establish of Scientific Panel as multidisciplinary, evidence based, and representative body that provide policy relevant scientific assessment on the risk, opportunities, and governance of the artificial intelligence, as well as impartial evidence based advice aimed at ensure fair and equitable, enabled and inclusive access to AI.
Throughout this process, we have consistently underscored the urgent need for the comprehensive international frameworks that guarantee the full inclusion of developing countries in shaping the future of AI governance.
In this regard, developing countries must also benefit from the impartial scientific guidance of the panel in order to foster technological autonomy, ize opportunities, manage risk, bridge capacity gap, and develop national policy that are knowledge and drivable and equitable.
The success of this effort will depend on a strengthened international cooperation, prioritizing capacity building, affordable and sustainable financing technology transfer, and fair inclusive and equitable data governance and access to data, thereby promoting shared prosperity and ensure that AI serve as a tool for sustainable development for all.
This effort should both accelerate innovation and help bridge and prevent the widening existing digital and development divides.
The group reaffirmed that the independent panel should align with existing US initiative through coordinated effort, ensuring coherence, avoiding duplication, supporting systematic capacity building, and promoting global AI governance alignment.
In this light above, the group extend the following question to the panel.
What concrete recommendation is the panel considering to ensure equitable access to a technologies and infrastructure and address the widening AI divide between developed and developing countries? Has the panel considered how can innovative finance ocean be operationalized unpredictable and accessible manner? What approach is the panel considering to facilitate technology transfer and knowledge sharing on AI? How can as related to intellectual property and access to data be addressed? How will be the panel incorporate diverse regional perspective and local context in this assessment while preserving impartiality and avoiding conflict of interest? How does the panel envisage its work feeding intergovernmental processes and UN negotiations? Looking ahead to the global dialogue on AI governance in July, where the panel will present its first report, could you share how the work is progressing toward that milestone? In particular, do you anticipate that the report will include concrete and actionable recommendations including those relevant for developing countries? The Group of G 77 and China looks forward to the panel's first report and recommendation as well as the continued engagement and reaffirmed its commitment to ensure a fair, inclusive, transparent, and equitable and development oriented approach to AI governance? Thank you.
There's many questions, and so you can provide a general answer, but you can also answer in writing if you want later on to the group that will be very useful for us.
Thank you.
Thank you very much.
Thank you also for being very pragmatic.
We plan to have the answers at the end, but I will see if we have so many answers.
Maybe I give you also the floor in between.
I thank the distinguished representative of Og on behalf of the Group of 77 and China.
I now give the floor to the distinguished representative of St.
Kitts and Neves on behalf of the Caribbean community.
Thank you, Madam President.
I have the honor of delivering this statement on behalf of the 14 member states of the Caribbean community, Caricum.
Caricum welcomes the establishment of the Independent International Scientific Panel on Artificial Intelligence as mandated by the General Assembly Resolution 79 slash 325, comprised of scientific experts from all regions around the world.
Caricum further welcomes the convening of this interactive dialogue with the co chairs of the panel, and we thank them both for the briefing just delivered.
In light of the rapid evolution and utilization of AI technologies, it is important for the United Nations system to advance international governance and to ensure that AI is used safely and sustainably.
It is also important to consider the intersection of AI and other global and international policy areas such as finance, education, health, and security.
Madam President, within Cicon in addition to the efforts of each of our individual members are deploying to manage and make the most of AI, a number of our regional organizations are also working on harnessing the power of artificial intelligence.
In 2024, UNESCO worked with over 20 countries in the region and over 1,000 institutions to develop the UNESCO Caribbean Artificial Intelligence policy roadmap.
Roadmap has aided in shaping AI for the unique needs and challenges faced by small island developing states, SIDS, and it is built on four pillars, culture and creativity, governance and transformation, education and upskilling and resiliency and sustainability.
These guiding principles ensure that Caribbean CDs are prepared to protect their citizens and to promote economic growth.
In December 2025, the Caribbean Artificial Intelligence Task Force convened under the Caribbean Telecommunications Union, published an interim report on AI policies and recommendations for the region.
This was guided by principles such as human rights, sustainable development, data sovereignty, and regional collaboration.
The report outlines a regional AI policy framework and connects it to critical sectors for the Caribbean, such as tourism and the development of infrastructure towards improved resilience to climate change.
A key recommendation from the task force inspired by our SIDS friends in the Pacific, discussing using AI for early warning dashboards and damage assessment drones to mitigate climate disasters, as well as developing a regional AI resilience hub to coordinate datasets, simulations, and training in collaboration with nearby universities and the Caribbean Disaster Emergency Management Agency.
These discussions from the report will continue at the first Caribbean Artificial Intelligence Forum in Trinidad and Tobago on the 23rd and 24th of July.
This is where policymakers, private sector, civil society, and other stakeholders from across the Caribbean and beyond will convene to work together to further advance regional AI governance.
This forum will, of course, complement the work of the upcoming global dialogue on artificial intelligence governance to be held in Geneva on the sixth and seventh July.
That will also examine the intersectional dimensions of AI implications, capacity building, and access, and the development of safe and secure AI systems on a and policy on a global level.
Madam President, the United Nations and its member states must reckon with the current crossroads we face in this digital age and must be prepared AI is reshaping the global public order, and we must ensure that it will benefit all of humanity and not just a privileged few.
In this regard, we warmly welcome the work being undertaken by the Independent International Scientific Panel on Artificial Intelligence.
Thank you.
I thank the distinguish representative of St.
Kit Sen Neves on behalf of the Caribbean community.
I now give the floor to the distinguish representative of Saudi Arabia on behalf of the Arab Group, followed by the European Union and Estonia.
Cedar R.
Madam President, Your Excellencies, I have the honor to deliver this statement on behalf of the Arab Group.
I would like to thank you, Madam President, for organizing this dialogue.
We also express our appreciation to the co chairs of the Independent International Scientific Panel on Artificial Intelligence, M Maria Isa and Professor Joshua Benjau.
We thank them for the valuable briefings and for the efforts in support of international dialogue on AI governance.
Madam President, the Arab Group believes that AI brings about promising opportunities to achieve sustainable development, to improve quality of life, enhance public services and support innovation and economic growth on the condition that these technologies are developed and used in a framework that respects the principles of the charter and ensures they are used for the service of human development and peace.
In this connection, the Arab Group reiterates the importance for AI global governance to be comprehensive, participatory and based on consensus among member states while taking into consideration the varying levels of development and national priorities.
This governance should ensure that digital gap does not go wider or existing inequalities get deeper.
The group stresses the following.
States should have the sovereign right to develop their national policies in relation to AI based on their national priorities and conditions.
Second, fair and equitable access to technologies, capabilities, infrastructure, computer resources, and data to enable developing countries to fully use the opportunities come with these technologies.
Third, enhance capacity building, knowledge, and technology transfer, support of scientific research and innovation in a manner that narrows the digital gap and enables and empowers youth, women, and entrepreneurs.
Fourth, develop the ethical and responsible frameworks for AI that enhance the reliability and transparency and accountability, and that respect the privacy and human rights while taking into consideration cultural specificities and national and societal values of each state.
Madam President, the Arab Group welcomes the current efforts to implement the outcomes of the Global Digital Compact and the establishment of the independent International Scientific Panel.
We believe that the success of this process required continued dialogue between the governments, the United Nations, private sector, academic institutions, and civil society.
While maintaining the crucial role of member states in any future arrangements or mechanisms on AI governance, the AR Group also stresses that AI should be a tool to enhance the implementation and the achievement of sustainable development and the 2030 agenda.
It should be used to address global challenges including in education, health, food security, water resource management, and climate action.
In conclusion, the ARP Group reiterates its support to international efforts to build a a fair, comprehensive, human centered AI global governance system based on international cooperation and effective multilateralism to ensure that the dividends of these technologies benefit all peoples and serve peace, prosperity, and sustainable development.
Thank you, Madam President.
I thank the Distinguished representative of Saudi Arabia, on behalf of the Arab Group.
I now give the floor to the distinguish representative of the European Union on behalf of the EU and its member states.
Thank you, Madam President.
Co chairs, dear colleagues.
I have the honor to deliver this statement on behalf of the European Union and its member states.
Let me first congratulate you as team co chairs on your appointment.
We look forward to your first report, and we thank you for taking the time to brief member states today.
Technology is not inherently good or bad.
It is the way it is used that can lead to positive or negative outcomes.
This is why AI systems must be designed, developed and deployed in ways that respect and promote dignity, equality, inclusivity, and human rights, as well as democratic principles and processes so that humanity can benefit from AI tools and ensure humans are protected from AI enabled abuses and violations.
To champion this approach, the EU will continue supporting effective multilateralism grounded in international law and the rules based international order with the UN at its core.
Dear co chairs, we would like to take this opportunity to ask you some questions, particularly about the upcoming report, your ways of working, and the relation between the panel and UN member states.
A more detailed list of questions will also be shared in writing.
On the substance of the reports and thank you already for the presentation you made, we know the time was very short as the panel was constituted in March and will present the report already in July.
Would you be able to share the primary conclusion so far in more details and what significant open questions do you want member states to be aware of? Secondly, AI capabilities are evolving faster than annual reporting cycles.
What mechanisms is the panel considering for interim advisories or rapid response assessments between annual reports? The modalities resolution, for instance, provides the possibility of thematic briefs to be produced as needed.
The EU would like to suggest such briefs to cover the following topics, recent developments in AI, AI and the environment, AI and Human Rights due diligence, AI and education, AI in the labor markets, the impact of AI on children, AI in cybersecurity, malicious use of AI by non state actors, and AI's positive contribution to development and human well being.
The positive impact of AI could also be included as a valuable subfoc in the other briefs.
Thirdly, how does the panel ensure that its scientific insights and work will be visible and available to all UN member states and stakeholders, regardless of the country's level of technological development.
A second cluster of question we wanted to ask you is on the independence and the scientific integrity of the panel.
The values the independent nature of the panel.
Its members serve in their personal capacities independently of any government, company, or institution.
This implies that appropriate safeguards should be in place to ensure that entities affiliated with panel members do not participate in or have access to the panel's internal deliberations.
What mechanisms does the panel use in practice to manage potential conflicts of interest, especially given that many members have affiliations with major tech companies, AI labs or national research institutes.
Secondly, how does the panel ensure transparency and avoid bias when it comes to informal consultations and interactions with external experts, while maintaining synergies with other relevant AI initiatives at all levels.
Finally, on this point, how does the panel adjudicate genuine scientific disagreements? You mentioned that you had some, of course, but especially how will those disagreements be reflected in the report? Another cluster of questions is about the representation and the AI divide.
Concretely, how does the panel capture evidence on AI impacts in developing countries where data is scarcer? How does the panel ensure that its scientific assessments take into account the views and research agendas from less resourced institutions, including those based in developing, remote or less developed geographical areas.
Finally, on the relationship between the panel and stakeholders, what is the panel's view of how member states and other stakeholders should use its findings and what risks of misinterpretation or misuse of those findings would concern you most? Last but not least, the panel is not a regulatory body and will not set rules, enforce standards or prescribe policy.
When findings clearly point towards a needed governance response, how will the panel communicate that imperative to member states without overstepping its mandate? I'll stop here.
Thank you again very much for your work.
You have the EU support.
Thank you very much, but this really shows the engagements and the importance of the topics are very valuable.
As some of you already mentioned before, there's the opportunity to hand in also the question in written form.
I know it's a lot of work also for the panel and we're all going to Geneva, but they will report back afterwards.
We have heard the last speakers for groups coming now to member states reminding everyone that speaking limits for member states are 3 minutes.
I would like to give the floor now to the Singers representative of Estonia, followed by El Salvador and Brazil.
Thank you, President.
Estonia aligns itself with a statement just made by the European Union.
Let me add few elements in my national capacity.
We thank Honorable Professor Joshua Bengio and Nobel Laureate Maria Ressa for their leadership of Independent International Scientific Panel on Artificial Intelligence.
We welcome the panel's efforts to provide independent, multidisciplinary and science based expertise in support of the first global dialogue on a governance to be held in Geneva on sixth and 7th of July this year.
Look forward with keen interest to a forthcoming report of Independent Scientific Panel, which is due to be published on 1st of July.
For Estonia, AI is not a future prospect.
It is already a practical tool for improving public services, strengthening economic competitiveness, and advancing sustainable development.
Today, more than 200 AI and machine learning use cases operational only across the Estonian public sector, ranging from citizen services and education to environmental monitoring, health care, and public administration.
Estonia makes AIU transparent by collecting and publishing information about public sector AI use cases, giving citizens insight into how AI supports public services and decisions that affect their daily lives.
This approach also helps government institutions avoid duplication, share knowledge, and strengthen collaboration across the public sector.
At the same time, we recognize that AI systems must be safe, secure, trustworthy, transparent, human centric, and consistent with international law.
Trust in digital and AI enabled services rest upon exactly these fundamental questions.
Are they secure, understandable, reliable, and genuinely useful to those they serve? Estonia's own experience as a digital state has demonstrated that such trust is hard earned through practical safeguards, transparency, and consistent quality of service, fostering trust and enabling innovation.
President, as a co chair of the inaugural global dialogue on a governance together with El Salvador, Estonia has witnessed the strong interest of member states and stakeholders in building a less fragmented, more interoperable, inclusive and truly global conversation on AI.
Throughout our preparations, we have consistently heard how vital it is to ground policy discussions in scientific evidence, technical expertise, and diverse perspectives and to arrive at result oriented solutions.
Looking ahead, we would be interested to hear from co chairs how the panel intends to engage with member states and wider scientific community after the dialogue.
I thank you.
I thank the distinguished representative of Estonia and I now give the floor to the distinguished representative of El Salvador.
Thank you very much, Madam President.
First, we would like to warmly recognize the work of the co presidents of science Epic Panel, Joshua and Maria for their experience and commitment.
They are turning the panel's mandate into a reality.
We would also like to thank all members of the panel for the important work that they've done in a very short amount of time.
As Joshua and other delegations mentioned, the establishment of the panel along with the global dialogue on AI governance is a key step To enhance the role of the UN to support approaches that are inclusive, evidence based, and cooperative when it comes to artificial intelligence and as co president of the global Dialogue, along with Estonia, El Salvador deems that the complementary relationship between these mechanisms is very important.
Everybody knows this, but it is important to reiterate it.
The panel has a unique role to provide scientific evidence as well as review and evaluation and the dialogue offers a space for member states and all interested parties, consider the policy implications as well as priorities and opportunities offered by the evidence.
In that sense, the work of the panel will feed into our discussions in Geneva, especially different thematic groups within the dialogue, and we'll look at opportunity and implications for AI and reducing AI gaps, strengthening AI to make sure that it is safe and reliable and promoting and respecting human rights.
We look forward with great interest to your report during the global dialogue, as well as the ensuing deliberations, which will be grounded in the findings and contributions of the panel.
We would also like to say some words of appreciation in terms of the constructive exchange with the coaches of the panel when it comes to the session.
To introduce the report, we've been exploring ways that we can integrate your experience into our thematic discussions as well as other parts of the program as appropriate.
This will contribute to ensuring that both mechanisms inform and mutually strengthen each other, fully respecting their different mandates and functions.
For El Salvador, this is precisely the value of the structure created by the GA to bring together science, policies, and multi stakeholder dialogue in a way that we can make the most of AI at the collective level inclusively and to benefit everybody.
We hope to keep the collaboration going and to make sure that the panel's work significantly contributes to the global dialogue and also broader efforts in the UN when it comes to AI governance.
Thank you very much.
I thank the distinguished representative of El Salvador.
I now give the floor to the distinguished representative of Brazil.
Thank you, Mr.
Chair.
I would like to thank the President of the General Assembly for convening this interactive dialogue and I'd like to welcome the insightful briefings by the co chairs of the Panel of Jicial Intelligence and please, I would like to ask the co chairs to extend to all panel members our appreciation for their tireless dedication in the latest month.
My delegation aligns itself with the statement made by Uruguay on behalf of g77 and would like to say that Brazil's position remains firm.
The UN is the most legitimate, inclusive, and appropriate forum for global discussions on AI.
We must work together to avoid institutional, political, and technical fragmentation, which would undermine our collective ability to leverage AI for development in democratic governance.
The United Nations should continue to play a central role in fostering consensus on and disseminating the fundamental principles and values that underpin AI governance.
One example is helping ensure that technical standards favor interoperability and broad access.
This is not just a preference, it is an economic necessity for developing nations.
In this period of multilateral commitment, Brazil actively supports the UN's landmark initiatives on AI, including the global dialogue and the global network of exchange in cooperation on AI capacity building.
We were pleased with the appointment of Professor Teresa Rudermre to the panel.
A leading expert on responsive AI, she embodies the broader approach that Brazil has consistently advocates for.
We believe the panel's reports with evidence based recommendation will play a fundamental role to inspire and guide the nationals debate worldwide.
Another key aspect is information integrity and the defense of democracy.
It stands as one of the most urgent priorities in the current global and national political context.
AI systems now actively mediate how information is created, ranked, and monetized for billions of citizens.
To safeguard our democratic institutions against algorithmic bias, deep fakes, and coordinate manipulation, we must leverage existing multilateral frameworks.
Mr.
Chair, Brazil stands ready to work with you and all member states to build an inclusive, fair and legally sound international framework that puts human dignity at its center.
Thank you very much.
I thank the distinguished representative of Brazil.
I now give the floor to the distinguished Representative of India.
India welcomes this opportunity to engage with the co chairs of the Independent International Scientific Panel on AI, Maria Rasa, and Joshua Bengio and appreciates the efforts undertaken thus far to operationalize the panel.
We align ourselves with the statement of G 77 in China delivered by Uruguay.
As we move forward, it's important that the work of the panel remains firmly anchored in the mandate agreed by member states in General Assembly Resolution 79 slash 325.
The terms of reference adopted by consensus reflect a carefully negotiated balance among diverse priorities and perspectives.
Preserving that balance will be essential to maintaining confidence in the process and ensuring broad ownership of the panel's work.
We therefore encourage strict adherence to the agreed mandate, scope, and modalities established by member states.
Second, the panel's work should maintain a balanced, comprehensive and development oriented approach to AI.
While risks, challenges, and potential harms associated with AI deserve careful and rigorous examination, these should not become the sole focus of the panel's work.
Resolution 79 slash 325 recognizes both the opportunities and the risks presented by AI.
For developing countries, AI holds significant potential to accelerate progress towards ESDGs, strengthen public service delivery, improve access to education and health care, enhance agricultural productivity, and support innovation led growth.
The panel should therefore devote equal attention to identifying pathways through which developing countries can harness the benefits of AI while addressing capacity, infrastructure, data, and skills gaps.
Third, it is essential to preserve the independent scientific and technical character of the panel.
Its credibility will depend on its ability to provide objective, evidence based, multidisciplinary and geographically representative assessments.
The panel should remain a source of impartial scientific expertise and avoid becoming a forum for political negotiation or policy prescription.
India looks forward to the panel making a meaningful contribution to international cooperation on AI and to advancing the vision articulated that the India AI Impact Summit 2026 held in New Delhi.
As the first global AI summit hosted in the Global South, the summit underscored the importance of ensuring that the benefits of AI are accessible to all countries and that developing countries are not merely consumers of AI technologies, but in fact, active participants in shaping the future of AI governance.
In this regard, the Independent International Scientific Panel can play an important role by providing the objective evidence based scientific foundation needed to support informed global discussions and to ensure that AI governance frameworks are inclusive, development oriented, and fully responsive to the needs and priorities of the global South.
Thank you.
Thank you, Chair.
I thank the distinguished representative of India.
I now give the floor to the distinguished representative of Romania.
Thank you, President, Professor Ressa and Professor Bengio.
What a sight.
I have to say that looking at the work of the panel for the past several months, I do believe that the panel is readier than member states are to look at different angles of the artificial intelligence, and I think that's very good news.
We do need this important work of the panel.
I very much agree with the description and the adjectives that modulate the work of the panel independent scientific, which is very much important.
I have to say that we do trust the work of the panel.
The intelligence that is coming out of that, but also the encouragements and the benchmarks that the panel would be creating.
I think we as member states, we need that very much.
As we're using in many conversations at the UN but also in the public, different buzzwords related to artificial intelligence.
I think we have to look at what really matters at the end of the day.
Um, how do we create the governance? How we make use of one of the best technologies ever developed, which continues to develop.
I think artificial intelligence will be much more developed in a week from now than it is today.
And that's a fact.
Governance is not a simple issue.
Of course, it's going to be very difficult to operate far beyond the siloed approach that we have now.
Different governments, different regions with different approaches to what governance should mean and what it should be.
Here, one of the key question is whether the baseline for governance is the minimal or maximal approach.
There are countries actually that do not benefit as much as they would like to from artificial intelligence, and many of us actually don't understand what artificial intelligence is about.
It's not only a search engine, it's much more than that.
For me, it would be very much important.
One of the key and the most important issues I think Professor Benja referred to is the search of truth.
Truth, I think should be at the core of any conversation.
I don't have too much time, but the truth, I think, at the end of the day, should be at the core of regulation.
Regulating not for the sake of preventing artificial intelligence becoming autonomous, agentic, which is in itself is a good objective.
But I think regulating to protect the truth, I think that's the most important probably aspect, and you have our full support.
Thank you very much.
I thank the distinguished representative of Romania.
I now give the floor to the distinguished representative of Poland.
Thank you, Mr.
President, Distinguished co chairs.
Paul, I would like to express sincere appreciation to the panel for its valuable work and thoughtful contributions to the global discussion on AI.
Moreover, that one of the members of the panel is a Polish citizen.
We're happy and proud of that.
Poland views that the development of the artificial intelligence as a defining element of the digital transformation and of course approach.
In a cruel principle, AI must be human centric, trustworthy, and aligned with the fundamental rights and democratic values.
It is both our vision and the practical framework guiding the design and governance of the AI system.
We believe that innovation and responsibility must go hand in hand.
AI should be transparent, secure, and accountable and must not reinforce discrimination or inequalities.
Human dignity, fairness, and privacy are not optional.
They are the foundations of the trustworthy AI.
Protecting human rights is essential to our approach.
We are committed to safeguarding personal data, promoting equal treatment as well as developing inclusive AI solutions that respond to the needs of all.
We believe that human rights must be protected both offline and online.
At the same time, Poland digital strategy places citizens and their quality of life at its core.
We see AI as a tool to improve access to public services, enhance efficiency, and strengthen transparency, building a state that is not only digital but also fair and accountable.
We are investing in the foundation of our responsible ecosystem, including data, infrastructure, skills, and cooperation between science, business, and public institutions.
Inclusive governance is essential to ensure that the AI reflects societal needs.
Education and public trust are equally important.
Trust must be built through transparency, digital literacy, and open dialogue enabling society to benefit from the AI while remaining resilient to its risks.
At the international level, Poland supports the rule based multi stakeholder approach with the United Nations playing as a crucial role.
Aligning ourselves with the statement delivered by the European Union in the context of international relations and the question that was asked, I would like to pose the following question to the distinguished co chairs.
Which AI risks does the panel consider most acute and underappreciated in current multilateral decisions and discussions.
Let me conclude by retraining our goal to foster innovation by protecting human rights and strengthening social cohesion.
Thank you.
I thank the distinguished representative of Poland.
I now give the floor to the distinguished representative of Latvia.
Thank you, Mr.
President, Distinguished, distinguished colleagues.
We thank the co chairs for their briefings and join others in congratulating them on their appointment.
Latvia alliance with a statement delivered by the European Union and wishes to add the following remarks in its national capacity.
Latvia welcomes the establishment of the Independent International Scientific Panel on Artificial Intelligence, and the first report of the panel is highly anticipated.
It is also highly necessary to promote scientific understanding of artificial intelligence.
It will provide a shared reference point for multi stakeholder discussions at the UN on the transformative nature of this technology, including during the inaugural global dialogue on AI governance.
When discussing AI, we are usually navigating between opportunities and risks.
We would be keen to hear if the panel has identified any opportunities, positive real life impacts that are currently underappreciated or underutilized.
On the other hand, which AI risks does the panel consider the most urgent and requiring the most immediate attention, and which risks are likely to be the most consequential over the long term? Distinguished co chairs, we would also be interested to learn regarding AI and information integrity.
In particular, will the report address the challenges posed by the generative AI to information integrity? The risks range from potential loss of cognitive skills due to over reliance on AI tools to issues of control over what information and knowledge large language models provide, as well as the risks of use and distortion of large language models by malign actors.
Finally, the last point on linguistic diversity.
Inclusive AI also means AI systems that support linguistic diversity.
Has the panel considered how to ensure that AI models serve the interests of underserved languages? It is crucial for the preservation of languages and cultures in the digital age, as well as for promoting broader technology uptake.
Thank you and all the best for the future work of the panel.
I thank the distinguished representative of Latvia, Excellencies, and dear colleagues.
I think we've received a good number of questions from the member states who have taken the floor, and I think it's appropriate at this stage to turn the floor back to the co chairs so that we receive some responses from them.
I now give the floor to Mr.
Yoshua Bengio to go first before we come to miss Maria Ressa.
I'm going to start by clarifying a question that was raised in various forms by Uruguay, the EU, and India about what we did and our mandate.
The question is whether the panel is producing recommendations, and the answer is, it doesn't.
Um, There's a good reason for this.
This is not its role.
The global dialogue is really going to look at the evidence that we bring forward, hopefully and bring the preferences and values and choices of different countries, but it's not our role.
Otherwise, our panel would be very quickly politicized.
And it would be difficult to stick to the facts and the evidence and remove bias.
So this is also related to the EU's question on how to communicate without overstepping our mandate.
This line is something that's been very much present in all of our work, and it was something the representative of India reminded this assembly, so we completely agree and are going in that direction.
I could do you want to take another one? Yeah.
First, it's wonderful to hear all the common ground that one nation can't do this alone, that multilaterlism is the route to go, that the UN should be the one and to pick up from what Joshua said, I think what you will see in the report the preliminary report, imagine we had roughly a month to argue this all out and the debate was from risks to opportunities.
I will say, I think one of the most interesting things about the gathering is that the global South is very well represented and dare I say the global majority has a very deep voice in this to look for the opportunities.
But at the same time, I think the scientists themselves know the dangers, the risks that are there.
For the Caribbean, it's Prime Minister Mottley was the one who said that literally information sovereignty is incredibly important, who owns the algorithms that shapes our citizens, which dovetails into the questions and the comments about human rights and democracy and well, our agency, you talked about it being good for being human centric.
Then I'll end with this one thing and I can give it back to you.
The question of the EU on the independence, and I think it was echoed by Latvia, how do we do this? We are very aware of all of the different I think everyone is interested in what the panel is going to come out with.
I think there's a distinction between interest and interference.
This panel works in a field where governments, companies, civil society, we all have enormous stakes and of course, they're watching us very closely.
I What we can tell you is that that is very healthy.
But interference is different and against it, we've built safeguards from the start, conflict of interest disclosures that are ongoing, not one time, confidentiality arrangements, clear and limited roles for the people who support us, strict limits on who can access our meetings and our documents.
Every member serves in their personal capacity, not as a representative of any government, any company, or any institution.
That's not a slogan.
It's a condition of sitting on the panel.
Where there's a concern about undue influence, we bring it up.
We're very transparent about it among the panel, and that's actually been the learning experience and a very humbling experience for a journalist.
It's been incredible.
The test we hold ourselves to is really simple.
The findings must remain the independent scientific judgment of us.
That's what we're aiming for.
I want to continue on this question, I think that was raised by the EU on how disagreements were handled.
It's a very important question.
And as Maria said, it's how could I say, it's a work of love.
That means listening to each other and, you know, agreeing to disagree in some cases, but seeking consensus.
And again, as Maria said at the beginning in her speech, sometimes that means, you know, talking about the things that there is a sufficient majority that agrees.
Now, one interesting question that we did ask ourselves in Madrid.
What if one of us disagree so much about one of the conclusions statements in the report that they could hold off the whole thing.
We created a process by which at the end of the writing period, each member of the panel could have a time to look at the whole thing and sign off, basically a go, no go, giving each panel member the possibility of being in disagreement and this would be stated at the end of the report, but it didn't happen.
We managed to negotiate and so on.
But this is always something that is possible and we keep that door open so no one can actually veto what the majority of the panel members think.
I want to also say a few words about a question that was brought by both Poland and Latvia about the risks, which ones are most acute, least appreciated, most urgent in the short term, in the long term.
Obviously, this would take a lot more time than we have today, and I encourage you to look at the report.
But I will, um will give you my opinion about more abstractly, where is our collective blind spot and it is coming from the under appreciation of the different possible scenarios of where the technology itself could go in the future.
The report talks about this and many other international reports have been talking about this, that Clearly, we can see the trends in AI growing capabilities, but no one has a crystal ball and it could stop its progress or it could continue at the current incredible rate.
And so really all countries should take stock of that uncertainty, and ideally, we should, as a global community, be prepared for all those scenarios.
So it is difficult because it's not in the current moment.
It is not something in our face, although recent events can give us some clues.
So it requires an exercise in imagination of where it could go.
What the consequences could be for all of the questions have been raised already from the geopolitics and the security to how we could benefit from this technology.
Yeah, I'll stop here, but obviously, you'll see much more in terms of answering these questions in the report itself.
I'm going to add three more points because there were so many that came through and I know we will wait for your written questions.
Um, To pick up from what Joshua said, do we recommend policies? Actually, that's your job.
We will give you definitions, we will give you what the scientists say, where it is right now, as far as we know.
Again, building on what Joshua said, there is so much we also don't know.
So we recommend, but you must create and implement policies that will keep us all safer than where we are today.
I think the second point is about um You mentioned human rights, information integrity, democracy.
All of these things fall under our ability to tell fact from fiction.
This is something that we've discussed.
But again, I think let me speak for myself on this.
You've heard me say at the UN General Assembly last year that we are in our information Armageddon.
This is the battle.
And if we lose that battle for facts for all of us to sit in the same room and agree on the facts, every single problem we will lose.
That includes climate, that includes every single other demand that requires a shared reality.
We've also discussed this a lot, and we hope that when you read the full report that you'll see, and then the last one, you asked about thematic briefs.
We have so much ambition in this, and I will tell you that every panel member has spent way too much, much more time, certainly more than we had expected, but we feel this moment and we hope you spend as much time on this as we did.
So some of the briefs that we've discussed AI and the environment, AI and child safety, AI governance instruments, and the evaluation of their effectiveness.
We've also talked about broader sectoral briefs, the applications of AI in space, quantum, legal and judicial systems, as well as financial markets.
This is an incredible panel in that.
Again, how often can we clash and fight and come out with something for you? We will continue doing that.
Um, can I add something to this? Because the question came from the EU and the question also raised the issue of the speed at which EI is advancing in comparison to the ability of our governments, or societies, our international institutions to adapt.
This is also one of the conclusions discussed in the report.
There is what has been called in the literature now an evidence gap.
In other words, Um, Scientific evaluation of impact of AI is going to lag.
It takes months, years for these studies to happen, for the papers to be written, to be published, and so the technology itself is not waiting for these publications to be deployed at scale.
And so, you know, the consequence for policymakers is tricky because they might not have all the evidence, and you will not find all the evidence you would like in our report simply because it doesn't exist yet.
We can see some clues, we can see, you know, some signs in some cases, but there's always going to be a delay between, you know, strong evidence and the impact in society.
And so there's a sense in which What governments need to do is to take stock of that uncertainty, accept that there's this gap, and sometimes take decisions that may be either too conservative and trying to prevent risks that don't unfold, or the other direction, of course, is not considering the risks for which we don't have enough evidence and then finding that we are in trouble.
Um, and of course, it gets into the policy discussions, but there's a rational basis for this, um in the world of, uh, risk management, which requires the kind of input we can provide, which is, what are the stakes? Some risks have very high stakes, others are more manageable, and that should guide the decisions.
Oh, Sosa.
Okay.
Sorry.
This is the last and I'll bounce off Joshua again on this one.
I think I speak to you not just as a member of the panel, but also as a victim, right? I was getting an average of 90 hate messages per hour.
Bottom up social media attacks followed by law fair.
My government is no longer trying to jail me now, so I love it.
But as a victim of this, I think this is to echo what Joshua said, this moment requires courage for you to evaluate the facts we will give you and to act in a way that will protect the citizens around the world that will protect our humanity, that will enable the opportunities to move forward.
You will see how nuanced this is in terms of how our panel thought it through, and I hope we hand it to you to act.
Courage.
Thank you.
I thank the co chairs for those responses and clarifications and I would like to open the floor again and now give the floor to the distinguished representative of Costa Rica, followed by the Islamic Republic of Iran and Algeia.
Thank you.
Co chair.
Costa Rica thanks the co chairs of the Independent International Scientific Panel on AI for their presentation and welcomes this interactive dialogue as a concrete step in implementing the commitments made in the Global Digital Compact.
Costa Rica is closely familiar with this process.
Together with Spain, it co facilitated the resolution that established the panel and the global dialogue.
We view this not as a finished architecture, but as the beginning of a structured, flexible, and evolving conversation regarding the principles that should guide the future governance of AI.
We are convinced that this global governance must be built upon rigorous, independent, and inclusive scientific evidence grounded in a sense of human purpose.
The panel must play a fundamental role in providing states with a common, objective, and scientifically rigorous basis for formulating public policies regarding the opportunities, risks, and impacts of AI.
However, its value will also depend on that evidence being relevant to all realities, especially those of developing countries.
At the national level, Costa Rica's AI strategy, 2024, 2027 reflects this vision.
We have conceived of AI as a tool to improve public services, health, education, productivity, sustainability, and public administration, always guided by an ethical compass, human dignity, human rights, equity inclusion, transparency, data protection, human oversight, and gender equality.
At the regional level, the dialogue on artificial intelligence led by Costa Rica a few weeks ago highlighted a central concern.
Latin America and the Caribbean must not be relegated to the role of mere users, markets, data sources, or testing grounds for technologies developed in other regions.
The region must actively participate in the development, regulation, and governance of these technologies.
To achieve this, it is essential to close gaps in infrastructure, connectivity, data, talent, and regulatory capacity.
Costa Rica encourages the panel to pay special attention to the differentiated impacts of AI.
The equitable distribution of its benefits, its social and economic impacts, the preservation of cultural and linguistic diversity, and the essential role of international cooperation on AI to prevent a new development gap among states.
AI must enhance human capacities, not subjugate them.
It must bridge divides, not codify them.
We invite the panel to adopt this as the horizon for present and future work.
It should be the guiding principle for inclusive evidence based global governance geared towards sustainable development and evidence based.
I thank you.
I thank the distinguished representative of Costa Rica.
I now give the floor to the distinguished representative of the Islamic Republic of Iran.
Rahman Ibrahim, Mr.
President, I align myself with a statement delivered by Representative of Uruguay on behalf of G 77 and China.
At the outset, I would like to congratulate the two distinguished co chairs for their leadership in guiding this important process, co chairs.
History has shown that the best solutions emerge from collective wisdom and genuine multilateralism.
We therefore hope that this independent group, emboding the aspirity of multilateral cooperation will help advance practical solutions for the benefit of humanity.
We look forward to receiving your report.
Mr.
President, artificial intelligence is one of the defining technologies of our time.
It has the potential to transfer education, healthcare, public services, economic growth, and sustainable development.
Yet, as we discuss the future of AI, one concern deserves particular attention, the growing AI divide between countries.
If left unaddressed, the AI divide risks becoming the next developing divide, countries with access to data, computing power, talent, and digital infrastructure will advance rapidly while other risk being left further behind.
This is why the question before us is not only how to advance AI, but how to ensure equitable access to the opportunities and benefits for developing countries, AI should be a catalyst for achieving development and sustainable development goals, not another source of inequality.
Bridging this gap requires international cooperation, solidarity, technology transfer, capacity building, knowledge sharing, and affordable access to digital infrastructure.
At the same time, AI governance must be inclusive, transparent, and representative.
The voices, priorities, and development needs of developing countries should be reflected in global discussion, decision making processes and reports.
We also underline the responsibility of technology companies, developers, and digital platforms.
Innovation should go hand in hand with accountability.
AI systems should be ethical, transfersive, free from harmful biases, protective of children, respective of cultural, religious and civilizational diversity and fully respect of national laws, regulations, frameworks, development priorities, and surveillance of a state.
And then we pose this question.
As global AI governance framework evolve, what create measures can ensure that developing countries are not receiver of the AI technology, but equal partners in shaping the rules and priorities of the global AI ecosystem.
I thank you.
I thank the distinguished representative of the Islamic Republic of Iran.
I now give the floor to the distinguished representative of Algeria.
My President.
Thank you, sir.
Professor Maria Ressa, Professor Joshua Begu.
I'd like to wish you a very warm welcome here.
Thank you very much for your briefings.
And for the initial responses you gave to our questions.
This is a very extraordinary opportunity to have this exchange with the co chairs on the Independent International Scientific Panel on artificial intelligence, a increasingly omnipresent technology that will play a growing role in all aspects of our daily lives.
This technology, with all its mysteries and uncertainties, isn't going anywhere.
It's here to stay because of the promises that it brings, but also the risks that it has.
It has already, stepped up and amplified longstanding challenges in particular disinformation, manipulation, inequalities, and other threats and our colleague from Colombia in particular highlighted this harms to human dignity.
These risks are particularly serious for the most vulnerable of countries.
The least developed, those from the global South, they are already facing digital, economic and technological divides.
Their ability to keep up with the pace of this technology poses existential questions of sovereignty resilience or rather essential questions of sovereignty resilience and society and institutions adapting.
Many experts are warning us about these artificial intelligence systems that are becoming too powerful and too autonomous.
We have to recognize that often All companies are acting with a logic of competition, power, and profit and it takes our governments much longer to take decisions in terms of containing this progress and dealing with it.
As you rightly said, Professor Ressa, we are very much looking forward to the panel's report, and now we are very, very much looking forward to looking at the fruit of what you are calling your specific disagreements.
However, and I'd like to put two questions here.
In your opinion, where should developing countries focus their efforts? They have limited resources so that they don't miss out on this AI train, so to speak.
The microphone is being cut off.
I thank the distinguished representative of Algeriia.
I now give the floor to the distinguished representative of France, followed by the Republic of Korea and Nepal.
Merci Monsieur President.
Thank you very much indeed.
Sir, co chairs of the Scientific panel.
Colleagues, I'd like to thank the co chairs for their presentations and all members of the panel for the enormous work carried out in a short period of time.
AI is a technological revolution and it talks to us about political, military, and economic powers.
This is significant every day.
It can bring both good and evil and it's up to our organization to contain the danger and to maximize the common good and this is why a full session of the G seven in Evian was on this a few days ago with companies there as well.
This led to the adoption of a statement on the protection of childhood in the digital space.
Co chairs.
There is a shared scientific understanding of AI to benefit all states equally.
France confirmed this in the 2025 summit.
This is a crucial part of AI global governance.
And we hope that the solutions that we will come up with together will be part of this consensus.
Your challenge is enormous, but so is the potential of AI to transform our societies, our economies, our children, and all of the challenges here.
Your entire complete independence is therefore an absolute prerequisite so that you can implement your work properly.
France, when we were just finding your roadmap, sought to ensure this and I want to stress this neutrality, impartiality, detachment from any particular interests, from any considerations of power and money.
Your independence needs to be upheld from all of these, um, standpoint, this is going to determine your legitimacy.
We want to encourage you not to spare any effort at all and not to be scared of sounding the warning bell.
The report that you're going to submit to us soon will clarify the UN's path regarding the AI and will clear the way for the AI summit in Geneva.
We'll talk there about the opportunities and risks that we need to think about in order to ensure that AI is safe, accessible to all and is for the public good.
I would like to finish my statement By urging you to ensure that you are heard, the world needs clarity on this very important issue of our time.
This is something that we see humanity wants to ensure that these voices of those who want this to be for all humanity be heard.
Thank you.
I thank the distinguished representative of France.
I now give the floor to the distinguished representative of the Republic of Korea, followed by Nepal and Indonesia.
Thank you, Mr.
President.
Warmly welcome co chairs.
I much appreciate for your providing much needed and promotive briefing and your insight.
Thank you.
At the outset, I would like to express the Republic of Korea's full support for the global dialogue on AI governance and the International Sn Panel on AI.
While the dialogue opens a space for exchange of the diverse perspectives, the panel serves as the scientific foundation that keeps the dialogue improved and helps bridge the knowledge gaps among the member state and stakeholders.
I'd like to offer three comments.
First, The panel slip should give us a clear mapping of where we currently stand in the development of AI.
The prop of the speed of the development and adoption of the AI.
The matters because, uh, AI governance must be evidence driven rather than peer driven.
So as the task is to govern AI in a way that neither st innovation no waste program to be materialized.
And only a clear scientific picture of the present can guide us toward the balance.
We look to the panel to establish with the scientific legal precisely where the case for global, not only the governance, but also global governance on AI Ls and why.
Second, apps to the Globan AI do not take place in a vacuum, nor do we need to reinvent the wheel where existing pymks already works.
So we hope the panel will help identify where existing pymks remain people purpose and where new challenges remain in made.
In the same speed would welcome the panels counted on the value and limitation of existing governance efforts and on the proper role of our global dialogues should play.
And lastly, we report the panels sharing the preliminary report with member state ahead of the post dialogue.
Today, you share the skeletons.
We need the precious.
Timing, I know the time is very challenging, but all delegations arriving grounded in a I thank the distinguished representative of the Republic of Korea.
I now give the floor to the distinguished representative of Nepal, followed by Indonesia and Italy.
Co chairs, Excellencies, Nepal thanks the co chairs for the comprehensive briefing.
We welcome the panel's progress and strong pursuit towards a global AI governance architecture that will shape the principles and transformative potential of AI for our people, societies and economies.
We further welcome the establishment of the independent International Scientific Panel on Artificial Intelligence as a multidisciplinary, evidence based, and representative body that provides policy relevant scientific assessments on the risks and opportunities and governance of artificial intelligence.
For Nepal, AI presents both opportunities and challenges.
We're optimistic about opportunities AI can unlock, including the creation of an entirely new economic sector and ecosystem as our vibrant and brilliant youth entrepreneurs Engineers, researchers in Nepal continue to explore, innovate, and create new potentials in AI.
But while algorithms increasingly reach our farmers, students, small businesses, and local governments, we continue to inherit systems without international mechanisms to assess their standards.
Challenges in transparency, auditability, and data governance creates risks of bias and uncertainty.
Uneven capacity, technological, and financial limitations and affordability further widen the gap.
In this context, first, the panel's work should focus on making AI governance inclusive and equitable.
Developing countries, particularly LDCs, landlocked developing countries, and mountainous countries like Nepal, should have meaningful participation in shaping global AI policies and standards.
Therefore, closing the digital and capacity gap is essential.
Second, AI's environmental footprint must become a core governance priority.
Global AI governance frameworks should set standards for energy efficiency and green AI practices.
Third, the panel's work should focus and promote AI governance that accelerates economic transformation.
For countries like Nepal, AI can support disaster risk reduction, climate adaptation, agricultural productivity, healthcare delivery, and education.
But this requires targeted investments for building digital infrastructure, digital public infrastructure, affordability, connectivity, large scale capacity building through regional centers of excellence, joint research platforms, practitioner level training, and strong networks linking startups, public institutions and technical communities within and between countries.
The panel can look into such ecosystem approach that enable countries not only to adopt AI, but also to build, regulate, and govern it effectively, scale up investment, strengthen their economy, support job creation, and close the digital divide we aspire to.
Closing, I have a question, how can the AI panel support governance frameworks that incentivize shifting AI infrastructure and AI compute infrastructure towards renewable energy to address the energy? I thank the distinguished representative of Nepal.
I now give the floor to the distinguished representative of Indonesia, followed by Italy and the Russian Federation.
Thank you.
Indonesia aligns itself with a statement delivered by Uruguay on behalf of G seven and China and wishes to add the following in its national capacity.
At the outset, Indonesia expressed its appreciation to the co chairs for the insightful briefing and also the members of the panel for their hard work.
In this regards, Indonesia would like to highlight three key points.
First, the panel should continue to serve as an independent, multidisciplinary, and evidence based body whose strength lies in bridging science and policy.
Second, Isia looks forward to the panel's first report.
We hope the report will highlight how AI can accelerate sustainable development in developing countries, as well as assessment of the enabling conditions required, including digital infrastructure, energy, data governance, talent, and most importantly, financing.
We also look forward to the identification of key evidence gaps on AI's impact on development to support more inclusive and informed global AI governance.
Third, Indonesia and courses the panel to ensure its assessment reflect evidence from all regions.
AI research evidence from developing countries remains limited.
Therefore, we hope the panel will engage global south researchers and expand regional case studies to build a balanced global evidence base.
Indonesia stands ready to contribute to this effort.
To close, allow me to pose a couple of questions.
The first one is, how does the panel address the intergenerational perspective on AI, given that different generations from baby boomers, Ganz millennials, Alpha often have different views and perspectives and approaches to AI.
Second question with the early warning nature of the panel, how you identify the most important question that science has not yet answered, especially those that matter most for the developing countries.
I like trends, risks or even emerging phenomena that others may often overlook.
Last, as AI evolve at an unprecedented pace, how will the panel keep its assessment timely? Thank you very much.
I thank the distinguished representative of Indonesia.
I now give the floor to the distinguished representative of Italy.
Thank you, Distinguished chair, distinguished co chairs, esteemed colleagues.
Italy aligns itself with the statement delivered on behalf of the European Union and its member states and wishes to make the following remarks in its national capacity.
Allow me to begin by expressing our sincere appreciation to the Independent International Scientific Panel on Artificial Intelligence for its valuable and timely work.
Italy remains firmly committed to supporting its mandate as a universal and inclusive forum to advance international cooperation grounded in a common understanding of AI and a shared scientific foundation.
AI represents an extraordinary opportunity for humankind.
It can foster economic growth, social well being, and sustainability.
However, we are convinced that its full potential can only be realized through a human centric approach that safeguards fundamental rights and upholds human dignity.
For these reasons, Italy supports, both in multilateral fora and its bilateral partnerships, the development of safe, secure, and trustworthy AI.
We advocate for the full implementation of the principles of responsibility and accountability, addressing the risks that the misuse of technology may pose to the protection and promotion of fundamental freedoms.
Innovation should not be regarded as an end in itself, but rather as a means to advance human well being, reduce inequalities, and strengthen democratic institutions.
Trust is therefore a prerequisite for the development and adoption of these digital technologies and such trust can only be built through transparency, accountability, and full respect for international law.
In this regard, we attach great importance to the role of the United Nations as the universal platform for dialogue on a shared approach to global AI governance.
Seeking common ground also means ensuring that the digital divide does not extend to AI.
Italy, therefore, considers capacity building a strategic priority as it is essential to ensure equitable and meaningful participation in the global digital transformation.
Investing in education, research, digital skill, infrastructure, institution knowledge sharing is key to enabling all countries to harness the opportunities offered by AI.
Through the AI for Sustainable Development, the digital flagship Italian Initiative, and its support for the UN university programs, Italy contributes to the establishment of resilient, inclusive and sustainable digital ecosystems, especially in Africa.
I thank you.
I thank the distinguished representative of Italy.
I now give the floor to the distinguished representative of the Russian Federation, followed by Canada and Morocco.
Thank you very much, distinguished colleagues.
We are grateful to the Secretariat and the co chairs of the Independent International Scientific Panel on holding today's meeting and for the substantive discussions on the preparation of the first report, we trust that the document will be submitted on time during the global dialogue on AI in Geneva, we believe that the report and the work of the panel should be fully in line with the mandate determined in the GEA Resolution 79-325 and pertain purely to the non military area of the application of AI.
We expect that the text will be professional, politically neutral in nature, and we'll note the scientific developments and will be marked by veracity.
We do believe that these will be fulfilled based on what you've said.
The panel's work should primarily be focused on ensuring the interests of the global majority, assisting the speeding up of the implementation of the SDGs and overcoming digital inequality.
We believe that this is one of the key challenges in the area of AI at the moment.
We're concerned by the current concentration of infrastructure, technology, resources and expertise in the hands of a small group of states.
These are essential for the development of AI.
There is currently a lack of full and equal access to this for the majority of developing countries.
We'd be grateful for comments from the co chairs as to the extent to which this issue is reflected in the report, and I mean inequality here and your considerations as to how to bridge this AI unequal divide.
Furthermore, it would be interesting to hear the view of the scientific panel as to what steps should be taken to ensure human oversight of artificial intelligence.
You said a lot that this is developing much more quickly than we think, in particular, given the active development of recursive self improvement, whereby the models themselves create data to ensure the existence of future generations.
So we wish you every success.
Thank you very much.
I thank the distinguished representative of the Russian Federation.
I now give the floor to the distinguished Representative of Canada.
Co president, Professor Bengio, Madam Ressa, delegates, I would like to first thank the co chairs on the panel for their leadership and their very hard work.
I would also like to thank the co chair of the scientific group of experts on AI, all 40 of them for their contributions to the drafting of this very important report that will be introduced during the dialogue.
Canada recognizes the importance of the work of the group of experts and their report.
It is based on evidence and rigorous scientific data.
We look forward to seeing it as a country that Thanks.
Responsible and human centered AI is very, very important.
We consider that this report will be an important contribution to strengthen our collective understanding and to really know what are the risks and opportunities to foster a shared scientific understanding of AI, recognizing that a common evidence based foundation is essential to building trust, supporting inclusive innovation, protecting human rights and the most vulnerable groups in our societies, as well as ensuring that the benefits of AI are accessible to all.
These commitments are reflected at the national level in Canada through Canada's recently adopted AI strategy, AI for all.
Building trust is a central pillar of our strategy.
It includes strengthening trusted international partnerships to advance AI research and promote responsible AI.
Responsible AI is safe AI.
Safe AI is based on trust, and trust is the basis of all of what we can do with respect to AI.
Expanding access and fostering inclusive innovation is another core component helping to ensure that AI uplifts workers, empowers businesses, and strengthens communities so benefits are widely shared, not concentrated amongst a few.
As we approach the review of the Global Digital Compact, we will remain actively engaged both within the United Nations and through complementary initiatives such as the International Network of AI institutes, the global Partnership on AI, and through our AI capacity building efforts to support secure, responsible and trustworthy AI.
With this said, we would welcome further insights from the co chairs of how member states can best engage with and benefit from the panel's work and how its findings will inform other multilateral processes.
We look forward to engaging further at the dialogue in July and continuing this important conversation.
Thank you.
I thank the distinguished Representative of Canada.
I now give the floor to the Distinguished Representative of Morocco, followed by Switzerland and Egypt.
Mercy.
Thank you, Chair.
First, my delegation would like to thank the presidency of the GA for organizing this interactive dialogue and both co chairs of the panel for their presentations.
AI, has become a major industry with a huge impact, but it's still concentrated.
It will reach billions of dollars in 2033, but only a handful of companies have all of the research.
Two countries have over 60% of patents and a minor amount of investment.
But most countries of the global South are not really players at the table where the rules are being defined.
Morocco addresses these technologies with clarity.
The opportunities are real.
AI can accelerate research, strengthen education, and improve public decision making, but the risks are there.
Disinformation and fake news make our democracies fragile and opaque systems can increase discrimination.
Data privacy is also a constant concern, and the labor market will be potentially affected in terms of 40 or 60% of jobs according to the IMF.
And the ideas to separate hope from fear.
This is why we need common scientific understandings.
We count on the panel to give us a solid basis to understand where are the opportunities and what risks we should foresee.
For Morocco, the end goal of these technologies is sustainable development.
AI will be valuable if it upholds Agenda 2030 and upholds our goals beyond that.
If it helps us detect disease before they strike, if it helps us better manage water and provide more just public services.
This is why we ask the panel to tell us what AI does for development and not just the ensuing risks.
And this promise means that conditions must change today.
Billions of people do not have access to the Internet and in Africa, only a third of the population is connected.
We see a new digital gap, the AI gap, but we don't really have data for a whole country.
This is going to be a problem.
The first report of the panel has to name this fracture, this gap and tell us how to close it.
Now we have the challenge of governance.
Over 80 countries are creating their own national rules, but national rules stop at borders and AI does not stop at borders.
We need common action and without it, the rules for this technology will be decided in closed circles, billions of people will be excluded.
This is why the panel is very important.
Until now, governments have knowledge about AI from companies that are creating in Geneva in a few weeks for the first time, each delegation, the microphone is cut off.
I thank the distinguished representative of Morocco.
I now give the floor to the distinguished representative of Switzerland, followed by Egypt and Ethiopia.
President.
President, allow me first to thank the co chairs for their work and thank them for their commitment to this process.
For Switzeralland, the creation of this panel is very important.
We consider that it is an essential element of the emerging architecture on global AI governance.
For this panel to really provide value added, its independence must be carefully preserved.
Its credibility depends, most of all, on excellence, diversity, and integrity when it comes to its scientific expertise.
Its role is not to draft political recommendations, nor to substitute intergovernmental processes.
Its mission is different, but it is also critical to provide states and other interested parties in a solid objective and accessible scientific basis.
The panel must translate the most advanced scientific and technical knowledge into language that is understandable and usable by public decision makers in an area that is characterized by extremely rapid technological change.
Bridging role between science and public policy will be crucial Switzerland believes this is very important.
Our country has a long tradition of scientific excellence and innovation.
We are convinced that effective public policies must be grounded in robust knowledge and a rigorous understanding of technological realities.
In this context, it is also critical that each component of the arrangement established by Resolution 79 slash 325 can fully carry out their specific mandate.
The global dialogue on AI governance and the scientific panel have complementary but different goals.
And this complementarity is an asset.
It allows us to avoid institutional duplications while enhancing the coherence of the entire system.
In an international landscape that has a plethora of initiatives, we must opt for synergies rather than fragmentation.
Finally, Switzerland hopes that the work of the panel will contribute to enhancing AI governance that is fact based, open to all regions of the world that pays attention to the needs of developing countries and is capable of enriching international discussions impartially and in an inclusive manner.
We look forward to the opportunity to continue exchanging and the conclusions of the first report that will be introduced at the global AI dialogue in Geneva.
I thank you.
I thank the distinguished representative of Switzerland.
I now give the floor to the Distinguished Representative of Egypt.
Thank you, Mr.
President.
Egypt would like first to align itself with the statements delivered on behalf of the group of 77 and China and the Arab Group and would like to add the following remarks in our national capacity.
We thank the co chairs of the Independent International Scientific Panel on Artificial Intelligence for their briefing and commend the work undertaken thus far in advancing the mandate entrusted to the panel by member states.
Egypt supported the establishment of the panel as an inclusive, multidisciplinary and evidence based mechanism, capable of providing impartial scientific assessments and policy relevant advice on the opportunities, risks, and governance dimensions of artificial intelligence.
The panel is a confirmation of the fundamental role that the UN should be taking in advancing the discussions on AI and its governance through dialogue between member states.
We also believe that this mechanism can make an important contribution to ensuring that the benefits of AI are harnessed in support of sustainable development and shared prosperity.
The same time, the emergence of artificial intelligence risks, deepening existing digital and development divides.
Persistent asymmetries in access to computing infrastructure, data resources, financial means, and technological capabilities could leave many developing countries at the margins of the AI revolution.
Mr.
President, in line of the above, we would be interested in hearing from the co chairs on several issues.
First, how does the panel intend to ensure that its assessments adequately reflect the priorities and realities of developing countries, particularly those relating to capacity building, infrastructure gaps, and access to data and computing resources.
Second, how does the panel envisage promoting stronger scientific cooperation and research partnerships between institutions in developed and developing countries with a view to narrowing the technological gap and fostering innovation capacities.
Third, given the importance of ensuring coherence within the UN system, how does the panel foresee its work complementing existing UN processes and informing intergovernmental deliberations while preserving the central role of member states in decision making on AI governance? I thank you, Mr.
Chair.
I thank the distinguished Representative of Egypt.
I now give the floor to the distinguished Representative of Ethiopia, followed by China and the United Kingdom.
Excellencies, Distinguished co chairs, Ethiopia aligns itself with the statement delivered by Urugay on behalf of the g77 in China, and we would like to add a brief remark in our national capacity.
Ethiopia welcomes the convening of this interactive dialogue by the co chairs of the Scientific Panel.
Let me also thank the co chairs of the Scientific Panel for their briefing and all the members of the panel for their commitment and dedication to this important process.
Ethiopia also extends its support to the scientific body and their crucial role in leading this process that will become a bridge between a scientific objective assessment and global AI governance and policy decisions.
Developing countries are experiencing the widening AI divide in various fronts, deepening already existing global inequalities.
We believe the upcoming report will become an important step in building a momentum towards effective international AI governance, widening the opportunities, managing the risks, and crucially bridging systemic capacity gaps.
AI is indeed one of the most consequential issues that the multilateral system must be equipped and prepared to address.
The UN's role is paramount as states and societies navigate this new, highly uncertain trajectory.
My delegation would like to pose a more methodological question.
I believe this is a basis for any scientific assessment.
In producing this scientifically grounded, objectively informed report, what systemic methodological challenges and gaps did the panel encounter and how these limitations were addressed.
In conclusion, Ethiopia looks forward to the first annual report by the scientific Panel and expresses its full support towards a more effective, sustainable and development oriented international AI governance.
I thank you chair.
I thank the distinguished representative of Ethiopia.
I now give the floor to the distinguished representative of China.
To President, thank you for convening this meeting and thank you to the co chairs of the Scientific Panel for their briefings.
China aligns itself with the statements made by Owai on behalf of G 77 and China.
China supports the UN in playing a leading role in global governance of AI and welcomes the establishment of the Scientific Panel and the global dialogue and the fruitful results achieved.
China supports the panel in upholding the principles of independence, objectivity, and professionalism, basing its work on scientific facts and latest developments in the field and serving as a bridge connecting science and policy and a link bridging the north and south divide.
We're glad to learn that the panel has been divided into seven working groups to conduct detailed research and analysis on topics such as cutting edge science on AI and AI governance.
We appreciate the efforts of the panel in producing its first reports in a short period of time, and we hope that the reports will bring more impetus to the development and governance of AI.
I'll make a few points.
First, respond to the concerns of the global Self on bridging the digital divide.
The world does not want to see an AI iron curtain and international cooperation on capacity building should be made a core agenda item to encourage the sharing of AI knowledge and the open sourcing of AI technology to promote and protect the diversity of civilizations.
Second, we hope that the recommendations in the reports be practical, effective, and actionable.
Building on attention to global AI cooperation initiatives, the report should identify best practices in international exchange and cooperation to provide countries, particularly developing countries with a scientific pathway and pragmatic recommendations for AI development.
Third, we expect the panel to continue to uphold the spirit of scientific neutrality and conduct objective research.
We hope that the panel remain open and transparent, actively interact with member states and stakeholders, and coordinate with global dialogue to form synergy.
President, China will continue to uphold genuine multilateralism, continue to deepen pragmatic cooperation in building AI capabilities, and work with all parties to promote the development of AI that is open, inclusive, universally beneficial, and for good.
I thank you.
I thank the distinguished representative of China.
I now give the floor to the United Kingdom, followed by Uganda and the United Arab Emirates.
President Co chairs, we thank the President of the General Assembly for organizing this session and to the co chairs for taking the time to brief member states today and to all panel members for the commitment to this work.
The UK welcomes the panel as an important source of independent scientific insight.
Its core value lies in strengthening the global scientific evidence based on AI to ensure that the international deliberations are informed by rigorous, independent scientific insight at a critical juncture for humanity.
The UK is a longstanding supporter of multilateralism and attaches great importance to strong scientific research.
We look forward to receiving the panel's report and commend the work undertaken to reach this important first step within a compressed time frame.
We encourage the panel to maintain a clear focus on core scientific questions, including adoption and diffusion dynamics and integration into critical sectors in a way that respects and maintains human rights.
We also support a focus on capacity building, particularly in strengthening access to scientific expertise, evaluation methods, and shared tools rather than infrastructure alone.
We would welcome the co chairs reflections on the following questions.
The impact of the report will depend on rigorous and transparent scientific methodology.
Could the co chairs share further detail on the methodology, whether it will be published, and how drafting has been undertaken between panelists and non panelists? How does the panel contend with disagreements? How will the report maintain synergies across existing processes, clearly delineate its scope and avoid duplication, including with technical safety reporting such as the International AI Safety Report.
How is the panel developing its work plan and how will it prioritize research questions across its three year mandate? As the panel will not prescribe policy or set standards, how will it convey when its findings point to a clear need for governance action in a way that respects its mandates.
Lastly, how is the panel structuring engagement with the global research community while preserving its independence? Could you speak more to the mechanism for wider stakeholders as mentioned in your statement? I thank you.
I thank the distinguished representative of the United Kingdom.
I now give the floor to Uganda.
Thank you, Mr.
President, Distinguished co chairs, miss Maria Ressa and doctor Joshua Benj, Excency's distinguished delegates.
We align ourselves with a statement delivered by Uruguay on behalf of the g77 and China.
We also wish to extend our profound gratitude to the co chairs for the insightful briefings.
We welcome your leadership of this historic panel, which was established under the GA Resolution 79 slash 325 as a concrete outcome of the Global Digital Compact.
This panel stands as a critical global early warning system and an evidence engine to bridge the growing technological divide.
For countries like mine, separating scientific fact from digital speculation is essential as we work to build robust national frameworks for AI.
As we anticipate the panel's report and the upcoming global dialogue on AI governance, we wish to underscore that it's very important for developing countries that it's highlighted that we lack access to baseline scientific infrastructure needed to evaluate advanced AI systems.
We urge the panel to actively level this information playing field by providing actionable accessible and independent assessments that enable all member states to anticipate risk and foster responsible innovation.
We underscore the urgent need for comprehensive international frameworks that guarantee full inclusion of developing countries in shaping the future of AI governance.
We also urge that the Global South must move beyond being a mere repository for raw data harvesting.
We request the panel to heavily investigate mechanisms for equitable benefit sharing and to prioritize sustainable development priorities such as leveraging AI for agriculture, healthcare, climate mobility, and localized educational resources.
The body must remain independent and must remain un swayed by competing national or commercial monopolies.
We support your focus on managing conflict of interest disclosures to ensure your output remains a universally trusted, reliable reference point.
As I conclude, we wish to ask the panel, how do you intend to constructively integrate and elevate local datasets and scientific expertise from African countries to ensure its upcoming assessments are fully global and globally representative.
You can counter our constructive engagement as we strive to ensure that the evolution of AI promotes the well being for all humanity.
I think.
I thank the distinguished representative of Uganda.
I now give the floor to the distinguished representative of the United Arab Emirates, followed by Pakistan and Kenya.
Mr.
President, thank you for coordinating this dialogue and it's truly a great privilege to hear from Professors Reesa and Bingo.
As you both highlighted, we are greatly in need of more neutral, accurate information and expertise on AI, which you and the fellow panelists are providing.
We are very proud that one MRT colleague, doctor Tukal Hanai is working with you on the panel.
The UAE aligns itself with the statement made by Uruguay on behalf of the G 77 and China and the statement by the Kingdom of Saudi Arabia on behalf of the League of Arab States.
We would like to raise just a small number of points which have not been covered yet.
The process you have outlined for the report drafting is highly impressive and the set of thematic clusters seems to cover well the most pertant issues for the future of AI and its governance.
Some priorities for the UAE at the upcoming global dialogue on AI governance will include how we can expand the opportunities of AI for development, how We can improve safety and responsible use of AI and the rights of women, children, and persons with disabilities in respect of AI.
We would be interested to know whether the report covers inclusion for those with disabilities, since this was not mentioned.
We also wanted to inquire whether the societal application section will include the deployment of AI and humanitarian work, including in the work of the UN itself.
Finally, we share the belief that the infrastructure around AI will be of essential importance.
Today, I would like to emphasize the implications of AI for energy, water, and the environment.
While AI is already exhibiting positive contributions to water and energy management, there seems to be a risk that the advancements of AI will impact progress on climate action and we would appreciate any further details on whether the report will address issues such as sustainable data centers, water recycling, and deployment of renewables in the powering of AI.
This is a topic that we hope will also be addressed at the UN Water Conference, which the UAE and Senegal will co host in Abu Dhabi in December.
Thank you once again and rest assured of the UAE's full support in reviewing, amplifying and exploring follow up action to the first panel.
Thank you.
I thank the distinguished representative of the United Arab Emirates.
I now give the floor to Pakistan, followed by Kenya and the Philippines.
Thank you, Mr.
President.
Pakistan aligns itself with the statement delivered by the G 77 and China and has the following remarks in its national capacity.
We thank the co chairs of the panel for this useful briefing.
We commend the panel for the important work it has been undertaking in a very short period of time.
The establishment of this panel is an important step towards ensuring that global AI governance is informed by credible, independent and policy relevant scientific evidence.
For developing countries, this work is particularly important.
Scientific guidance on AI must help countries build technological capacity, manage risks, develop knowledge driven national policies, and avoid a further widening of the digital and development divides.
At the same time, as highlighted by several member states, the panel's credibility will depend on its independence, transparency, and methodological rigor.
The panel was designed to operate with scientific independence, including from governments and the private sector interests, while contributing to the broader AI governance.
In particular, given the significant role of large tech companies in AI research and infrastructure, appropriate safeguards will be important to ensure that the panel's work remains balanced, impartial, and focused on broader societal and developmental impacts.
Against this backdrop, Pakistan would like to highlight following key areas.
First, the panel's credibility will depend on maintaining scientific independence while ensuring that its assessments remain policy relevant, practical, and responsive to the needs of member states.
Second, the development dimension of AI should be fully reflected in the panel's work.
This includes identifying the practical constraints developing countries face in relation to data, compute capacity, digital infrastructure, skills, financing, and technology partnerships, including accounting for localized contexts in the assessments, ensuring that global AIF frameworks natively support linguistic diversity and respect regional data governance.
Third, it was useful to hear about some of the thematic briefs that the panel is working on.
Given the compressed timeline for the first report, early clarity on the likely scope and areas of focus of the preliminary report to be presented at the dialogue would help member states engage more substantively in the process.
We hope that the panel's first report will provide a strong evidentiary basis for an inclusive and development oriented discussion at the dialogue.
Thank you.
I thank the distinguished representative of Pakistan.
I now give the floor to the distinguished representative of Kenya, followed by the Philippines.
Mr.
President, Kenya aligns with the statement of g77 and China and wishes to add a country specific statement on a subject that is rapidly shaping our economies, societies, and features.
Kenya commends the panel's work and believe that science, evidence, and inclusivity must remain at the core of global discussion of AI governance.
While AI present enormous opportunities for accelerating progress across sectors such as healthcare, education, agriculture, financial inclusion, and public service delivery, we must also be candid about the risks.
From our perspective, AI is not only a technical issue, it is increasingly a social issue.
We are already seeing concerns around bias, misinformation, exclusion, privacy, and potential displacement of jobs.
If left unaddressed, this risk could deepen existing inequalities both within countries and between developed and developing economies.
We therefore see a strong need for greater transparency and accountability in AI systems.
As models become more powerful and complex, questions of auditability, explainability, and independent oversight become increasingly important.
Policymakers, regulators, and public need confidence that AI systems can be assessed, tested, and trusted.
At the same time, discussions on AI governance must address the growing gap in access and capabilities.
Many countries in the global South remain on the margins of the AI ecosystem with limited access to computing infrastructure, data resources, research capacity, and technical expertise.
If these gaps persist, AI could become another source of global inequalities rather than a tool for sustainable development.
We would therefore encourage the panel to pay particular attention to the practical impacts of AI across different use cases and context.
The opportunities and risks associated with AI in health care may be very different from those in education, agriculture, financial services, employment or public administration.
For Kenya, the central question is how we can ensure that AI remains human centered, inclusive, and development oriented, advancing innovation while safeguarding trust, equity, and shared prosperity.
We look forward to the panel's recommendation and stand ready for contribution to this important work.
I thank you, Mr.
President.
I thank the distinguished representative of Kenya.
I now give the floor to the distinguished representative of the Philippines, followed by Bangladesh and Luxemburg.
Thank you.
Thank you very much, Mr.
President and we warmly thank the co chairs of the AI scientific Panel, Madame Ressa from the Philippines and Professor Bengio for their important work and presence with us today.
The Philippines aligns with the statement of the G 77 in China and provides the following points in its national capacity.
The Philippines was actively involved in negotiations on both the Global Digital Compact and the WSIS plus 20 review outcome document, which remain the UN's main documents that touch upon AI.
We highlight that AI governance cannot be the preserve of those who build the technology alone.
It must include developing countries seeking to harness AI's benefits for sustainable development.
We therefore see the work of the independent Scientific Panel through the UN as a key mechanism in this regard that can help contribute to creating conditions for a more inclusive, equitable and development oriented AI landscape and governance system with much more meaningful participation of developing countries.
Some priorities of the Philippines in this respect are first, bridging the digital and AI divides and developing more effective capacity building programs, especially in AI education and implications for climate action and disaster risk reduction.
Secondly, norm setting with human oversight over AI and digital technologies ensured that will help mitigate growing risks and disinformation among others.
Thirdly, addressing, as other colleagues have mentioned this morning, addressing the impacts of AI on labor and employment.
We would therefore be very interested to hear how the panel is looking at and addressing these issues.
We would also like to ask the panel's initial views and recommendations on how to strengthen international cooperation and also improve engagement with private sector actors.
We look forward to the presentation of the panel's report at the AI dialogue in Geneva in July and would also recommend, as others have mentioned, also developing specific science based and policy relevant recommendations for developing countries across the issues you're looking at.
Lastly, we would also recommend regular interaction between the panel and member states through formats such as this so that we could essentially work together as you develop your further follow up work.
Thank you.
I thank the distinguished representative of the Philippines.
I now give the floor to the distinguished representative of Bangladesh.
Thank you, Mr.
President.
Bangladesh allied with the statement delivered by Uruguay on behalf of G 77 and China.
Bangladesh thanks the panel for its valuable work and appreciates this effort to the wise discussions.
For countries like Bangladesh, AI presents major opportunities in many sectors, including health, education, agriculture, climate adaptation, and public service delivery.
However, these opportunities will remain limited unless developing countries have meaningful access to compute data, skills, open resources, and technology transfer.
Bangladesh also believes that the upcoming AI governance will address its impacts on labor intensive economies.
AI enabled automation may affect employment in sectors such as freelancing, IT and level services, governments, agriculture, and business process outsourcing.
In this context, I have two questions.
The first, how does the panel plan to address the risk of AI driven job displacement in labor intensive developing countries while ensuring reskilling, decent work, and inclusive economic transition.
And the related second question is, for countries that are more users than developers of AI, how can global governance and the panel's recommendations prevent dependency on the large tech companies from being leveraged against developing countries in the future.
We look forward to the valuable work of the Independent Scientific Panel.
I thank you.
I thank the distinguished representative of Bangladesh.
I now give the floor to the distinguished representative of Luxembourg.
Merci.
Thank you very much, Mr.
President.
Luxembourg endorses the statement that was made by the European Union, and I would like to take this opportunity to thank the President of the General Assembly for having given us this opportunity today to have an exchange of views with the Independent Panel on Artificial Intelligence.
I very much would like to thank the two co chairs, miss Maria Ressa and Mr.
Joshua Benjiu for their briefings today.
The outlines of the report that will be published on the 1st of July and that will fuel the discussions of the global dialogue on artificial intelligence in Geneva, a dialogue that our Prime Minister will participate in in order to underscore the importance that Luxembourg attributes to this.
I'd like to take this opportunity to put a question to the two co chairs, please.
In your presentation, you mentioned the main topics considered by the panel.
I'd like to ask you, please, could you give us more details about the panel's observations regarding the crucial need to ensure the protections of the right of children and also the need to ensure safety of childhood in the context of the developing of AI? Thank you very much.
I thank the distinguished representative of Luxenburg.
I now give the floor to the observer of the observer state of the Holy See.
Thank you, Mr.
Chair.
The delegation of the Holy Sea welcomes the convening of this first interactive dialogue with the co chairs of the Independent International Scientific Panel on Artificial Intelligence and wishes to thank both miss Ressa and Professor Bengio for their briefings.
As Pope Leo X XIV has emphasized in his recent encyclical letter, Magnifico Manits, new technologies are not yet fully predictable.
This complicates the assessment of their potential impact and the long term effects that they have on both the dignity of individuals and the common good.
In light of this consideration, the Holy see welcomes the mission entrusted to the scientific panel to issue evidence based scientific assessments and to evaluate carefully the risks and impacts of AI in order to ensure a sound scientific foundation for its governance.
In less than one month, the Scientific panel will present its first annual report at the global Dialogue on Artificial Intelligence Governance in Geneva.
It is our hope that this report will help inform the international community's discussions at a critical moment.
To that respect, a rigorous assessment of both the opportunities and risks associated with AI should be not seen as an obstacle to progress, but as an act of responsibility.
This is especially important given the widening gap between the rapid pace of technological change and the slower development of national and international norms, institutional frameworks, and safeguards needed to address its implication.
Mr.
Chair, technology is never neutral.
It is inevitable, inevitably reflects the priorities and assumptions of those who design, finance, regulate, and use it.
Every system embodies choices through what it measures, ignores, and optimizes.
For this reason, ethical considerations must inform the technical dimension from the outset rather than being treated as an afterthought.
In fact, as Pope Leo enlighted, ethical discernment cannot be limited to asking whether we are using a system for good or bad purposes.
It must also examine how that system is designed and what vision of the human person and the society is embedded in the data and models that guide it.
The Holy remains committed to engaging constructively in international efforts to ensure that artificial intelligence promotes human dignity, preserves the God given vocation of the human person, and upholds the common good.
Thank you, Mr.
President.
I thank the distinguished representative of the Holy See.
Excellency, distinguished colleagues.
We have had the last speaker and quite a number of questions and requests for clarifications have been raised.
At this point, I'll now turn the floor to the co chairs to try in the remaining time to provide the responses they're able to do before we can continue.
I'll start with Mr.
Benjau.
You have the floor.
Okay.
Thank you.
I will start with a topic that has been touched by several of you, Russia, Indonesia, the Holy Sea, Algeria, and France, that has to do with autonomy of the AI systems that are being developed and um how this relates to the dual use nature of these systems.
This is a topic that is discussed in our preliminary report.
I do think that there will be a lot more downstream as we see the related question of AI agents, you may have heard about, which is quite a bit discussed in our report.
AI agents are systems that can take decisions for themselves, and it's connected to another topic I'll talk about, which is labor impact.
Clearly, something that many companies are looking forward to is automating many tasks that are currently done by people.
Right now, we're talking about tasks that are happening through a computer, not physical tasks, but it may come also in the future.
To automate those tasks, which can increase productivity, but also create all kinds of impacts on society and workers, the AI systems need to be fairly autonomous.
You cannot have human oversight every 5 minutes on the work done by NAI, if it is going to be efficient in automating some of those tasks.
The two things go together, the benefits to productivity, but also the risks of the divide in terms of economic development that has been raised and the risks that we don't fully control those systems because they are taking their own decisions without a human behind their back all the time.
Um, so that connects to a question raised by Russia about, uh, recursive self improvement, which is sort of the extreme version of autonomy here.
This is also a question that is raised in some form in the report very briefly and deserves more.
It's a maybe longer term question, but clearly, it is the plan of the leading AI companies to build AI systems that will design their own successor that will do AI research and that could accelerate further the already, kind of, uh, Concerning speed of advances in AI, especially if we don't have full control over these systems, over the goals that they're trying to achieve.
Currently, this is not a question that science has an answer to.
Indonesia asked this question, what is the most important question science hasn't figured out about AI? Of course, there are many, but related to my discussion about autonomy, this is one of the questions.
How do we make sure AIs will behave according to our instructions, and it's not such a big deal when it's just a chat bot because there's a human in the loop all the time.
But it becomes a big deal with those agents that already do a lot of work by themselves, but it's going to continue.
Going back to self improvement, the facts about this are the speed at which the research in AI itself is being automated.
In several companies, we can see the proportion of say lines of code in developing AI.
Is written by AI and that proportion has been rising very steadily over the last year.
In fact, it's a sea change from last year today, and we continue to see this.
Yes, I'll have other things to talk about, but I'm going to pass it to Maria.
Well, ping pong.
First, fantastic.
It's wonderful to see where everyone agrees on certain things.
I'll also pick up on Indonesia and dovetail childhood and intergenerational.
These are topics intergenerational on its own is not necessarily specifically picked up, but we do come out with, and this is probably the one line I will read through because you should read the entire report in full context, but this is in relation to the types of methodologies that we could have.
Evidence from digital literacy programs indicates that AI literacy should be tailored to age groups, education levels, and cultural contexts.
I think that different countries have raised that.
On the question of both linguistic and an equitable access, which several of you have brought up, potential adoption, um, There was significant attention by the panel that was spent on languages in this preliminary report.
The fact that AI leaves most languages behind with specific examples and evidence that highlights the importance of adapting to the local languages, we're building our own in all domains and that includes some of the topics that you had asked about which are covered will include agriculture, education, culture.
There are case studies that the panel worked very hard on.
Then I think let me address this the Holy See.
Thank you for Pope Leo's cyclical.
It was something we discussed, something that we read and looking forward to more.
I In terms of how we work, and this is to address some of the I think Pakistan also you mentioned this, how do we maintain independence and how do we make sure that we actually give you something that isn't corrupted is the wrong word.
It's too hard.
It's a journalist word.
Sorry panel.
I But I'll talk about four things.
First, the membership itself.
We have been selected for disciplinary, gender, and geographic balance.
I have not seen something like this and many of the panel members are like me.
We're hybrid.
We live in one country, but we move like you in different places.
The second, we've said this, we serve as individuals, not as delegations.
Third, I think Joshua talked a little bit about this, but let me add a little bit more.
I Our drafting is collective.
It's part of the reason it's so time consuming.
Any member can challenge language, add evidence, or register that they can't support a formulation, and that dissent stays visible.
What we are also learning through this process is how we disagree in respectful manners, but disagree with courage and find common ground.
I really dislike that word though, we find that.
Then fourth, this is ongoing and not a box that we just tick.
We know that our legitimacy depends on continued engagement with every region.
I come from a country where we want to make sure that we're not left behind, but we also want to make sure that we minimize the harms.
This depends on continued engagement with every region, especially the developing countries and underrepresented communities.
We're doing our best.
I'm going to continue on the topic I started earlier.
I Going to a question from Indonesia, again about the questions that science doesn't have an answer for.
I talked about control of these systems.
There's a very concrete current issue in the minds of many people in the world of AI, but also cybersecurity, which is that the advances in capabilities that I've talked about are now past a threshold that allows the most advanced systems to be used by non experts to launch cyberattacks.
These cyberattacks could potentially be launched against software infrastructure that are now deployed across the world in ways that make all of us more vulnerable.
Of course, this has been a topic of a lot of discussion and is something to keep in mind.
Not just that, there are systems that now have these capabilities that are in some ways stronger and faster than humans are finding vulnerabilities in code.
But very likely if the capabilities of AI continue to rise, we'll see more of these kinds of things where the knowledge, again, an example of dual use, the same knowledge that allows AI to be used to do Research and programming, as I was mentioning earlier, can be used in a way that is disruptive and connecting to the question of Russia on the military applications of AI, the issue, the issues that are really about military application were not covered.
It is clearly in our mandate to avoid this.
But, in the case of cyber and other uses of the knowledge that the AI has, this raises security issues that are not just the military level, but you can think of criminal use or non state actors, terrorism and so on.
So this is clearly something you'll see more about in the report.
So I'll stop here.
Then just I'll pick up more on the child safety.
A quick summary because there are many sections.
We have several members of the panel who work on child safety, on human rights.
The panelists gave particular focus on human rights.
It literally cuts across the sections, more specifically on children's rights underneath, acknowledging that system level changes will create both opportunities and challenges.
The report talks about these risks that require special protection, such as sexual exploitation and abuse in health, well being.
There's another section I know someone had asked about agriculture, that's important to some of our panel members as well.
You will find that there.
Stay tuned.
Read it.
Great.
I thank the two co chairs.
I would like to thank Mr.
Bengio and miss Ressa for those responses.
Excellencies, distinguished colleagues.
I would like to thank all the delegations for their participation at this meeting.
And since we've had the last would you like to comment? One more.
Not at all, please.
So sorry.
On the UK, the AI safety report, Yasa chaired that one is one of the very significant sources of scientific knowledge and consensus on the question of AI capabilities and risks.
If you haven't seen it, please take a look.
It's been an important source for our panel.
We've taken it into consideration and relied upon the synthesis of the impacts and risks of AI.
You'll see it cited in our preliminary report.
That's our sneak preview.
I thank you once again.
Excellencies, colleagues, the informal meeting of the plenary is now concluded.
The meeting is adjourned.