EU Statement – UN General Assembly: Interactive Dialogue with Co-Chairs of the Panel on Artificial Intelligence

Dear Co-Chairs, dear colleagues,

I have the honour to deliver this statement on behalf of the European Union and its Member States.

Let me first congratulate you, esteemed 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. 

On the report

• The Panel was constituted only in March 2026 and will present a report in July. What are its primary conclusions so far, and what significant open questions does it want Member States to be aware of? 

• The Panel's mandate is to synthesize and analyse existing research, not produce original research. What significant gaps in the existing evidence base are most consequential for the report's findings?

• Which AI risks does the Panel consider most acute and underappreciated in current multilateral discussions? 

• Many reports on AI risks and opportunities focus exclusively on LLMs and AGI. Will the Panel pay equal attention to non-language-based models, embodied AI and industrial AI?

• In what areas does the panel see the biggest progress and potential in terms of applying AI systems?

• 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 for 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 and the labour market; the impact of AI on children; AI and 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 sub-focus in the other suggested briefs. 

• How does the Panel ensure that its scientific insight and work will be visible and available to all UN Member States and relevant stakeholders, regardless of a country’s level of technological development? 

On Independence and Scientific Integrity

• The EU highly 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?

• 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 (including at national, regional and international / UN levels)?

• How does the Panel adjudicate genuine scientific disagreements among members and how will those disagreements be reflected in the report?

On Representation and the AI Divide

• Concretely, how does the Panel capture evidence on AI impacts in developing countries, where data is scarcer?

• Has the panel's research examined potential regional variations in how AI is deployed, used, and diffused? And based on the available scientific evidence, would the panel say there are observable differences in AI's effects across different regions? 

• Has the panel also look at different implications for different socio-economic groups?

• 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?

On the Panel–Multistakeholder Relationship

• What is the Panel's view of how MS and other stakeholders should use its findings, and what risks of misinterpretation or misuse of those findings would concern you most?

• The Panel is not a regulatory body and will not set rules, enforce standards, or prescribe policy. When findings clearly point toward a needed governance response, how will the Panel communicate that imperative to Member States without overstepping its mandate?

• The Panel’s report will be essential to inform the conversations during the Global Dialogue on AI Governance in two weeks – by creating a shared understanding among all stakeholders through credible, independent research and evidence linked to AI risks and opportunities. How does the Panel envisage its interaction with the Global Dialogue on AI Governance, in particular in ensuring that its findings effectively inform and enrich multilateral discussions?  

• Does the panel plan any interim communication to stakeholders on progress and main assessments in addition to the presentation of its annual report to the Global Dialogue on AI Governance (e.g. through information events by its members)?

Thank you again for your work, you have the EU’s support.