GYA Alumna Maral Dadvar (Frankfurt Goethe University, Germany) joined an expert roundtable of the UN Scientific Advisory Board (UN SAB) to explore the challenges and opportunities surrounding Artificial Intelligence (AI) verification. The session was part of the Board’s ongoing work to produce a science brief for the UN Secretary-General, focused on how verification mechanisms can support the safe and equitable development of advanced AI technologies. The roundtable brought together a diverse group of experts in science, technology, and policy to provide insights into both current developments and anticipated future trends in AI governance.
Maral was nominated by the GYA to participate as one of the contributing experts. Ahead of the session, experts were encouraged to submit short written inputs. Maral’s written input – “Inclusive AI: The Need for a Global Strategy to Prevent Widening Inequality” – added an important perspective on accessibility and equity in global AI development, complementing the roundtable’s broader focus on technical and regulatory dimensions of verification.
The roundtable explored the definition of verification, with general agreement around the idea of enabling independent parties to credibly assess the claims of AI developers. Discussions touched on a range of approaches, from institutional audits and software-based transparency mechanisms to hardware-level verification and protections for internal whistleblowers. The group also debated the feasibility of compute governance, a concept gaining traction as a potential method for monitoring the use and deployment of high-performance AI systems.
Technological advancements were reviewed, including developments in auditability, system evaluation, and traceability. Several participants reflected on the evolving regulatory landscape, citing examples from the EU, the U.S., and China. These frameworks illustrate the increasing role of states and regional bodies in shaping AI oversight. In this context, many attendees expressed support for a stronger role for the UN as a neutral platform to encourage dialogue, promote harmonized standards, and coordinate verification efforts globally.
While the roundtable focused heavily on verification as a technical and governance issue, there was general agreement that efforts to improve trust and transparency in AI should also consider global inclusion. This perspective added depth to the conversation, reinforcing the importance of aligning verification with broader principles of equity and fairness.
Download UN SAB Policy Brief
The UN SAB Policy Brief can be downloaded here: Verification of Frontier AI Models | Secretary-General’s Scientific Advisory Board