AI Safety Connect convenes at UN General Assembly

AI Safety Connect (AISC) recently convened at the 80th United Nations General Assembly (UNGA) in New York City Sept. 25.

AISC is organized by the AI Governance Coordination Project (AIGCP). This ‘at UNGA 2025’ edition was co-organized with The Future Society and FAR.AI, and co-hosted by the Permanent Missions to the United Nations from Singapore, Canada, and Brazil, the United Nations Development Programme (UNDP), and Mila.

What began as an initiative to accentuate AI Safety during the AI Action summit in Paris now brought together scientists, frontier labs, and policymakers during the UN high-level week, which AISC said signals AI safety is moving onto the same level of urgency as climate change and nuclear security.

Stakeholders gather to discuss AI governance

AI Safety Connect said it is where the world meets to make AI safe. Launched in early 2025 during the Paris AI Action Summit, in just a few months it has evolved into an international platform for AI safety and governance. AISC provides a neutral convening space where global stakeholders can address the risks of an unchecked race toward superintelligence.

Every quarter brings new AI breakthroughs. Systems are becoming more powerful, agentic, opaque, and difficult to control, which AISC said exacerbates the risk of their unintended consequences and deliberate misuse. No one frontier lab or country can contain the promise or mitigate the harms and risks posed by advanced AI systems.

Harnessing global and multilateral coordination requires deep, trusted, and consistent engagement with stakeholders. AI Safety Connect said it provides a place to unite stakeholders on pressing global AI safety issues and discuss practical mechanisms for governance that enable responsible innovation and sustainable development.

By embedding AI safety into multilateral discussions at the UN, the initiative ensures that governance frameworks evolve in step with technology and that cooperation does not stop at borders.

“AI safety is no longer a technical issue - it is a diplomatic one,” said Cyrus Hodes, AI Safety Connect co-founder. “With AI systems advancing at unprecedented speed, we need the same kind of international cooperation that has governed nuclear security or climate change. AI Safety Connect exists to make sure this conversation happens at the highest level and continues beyond single events.”

Gathering called for acceptable use definitions

Around the 80th UNGA, AI Safety Connect convened roughly 100 high-level participants from UN agencies, governments, frontier labs, academia, and civil society to:

• Coordinate policymakers’ responses to advanced AI development by frontier labs across borders and industries.
• Highlight how advanced AI affects the role of AI safety institutes, and government more broadly.
• Introduce a Global Call for AI Red Lines, a campaign calling for international action to define unacceptable uses and behaviors of AI systems. As of today, it has been endorsed by over 200 figures, including former heads of state and Nobel laureates, and over 70 organizations working on AI governance.

“Our mission with AI Safety Connect is to build lasting channels of trust and cooperation between policymakers, researchers, and frontier labs,” said Nicolas Miailhe, AI Safety Connect co-founder. “This is not about a one-day forum, but about creating the infrastructure for collective action so that when AI reaches dangerous thresholds, the world is prepared to respond together.”