OpenAI CEO Sam Altman recently published a proposal in the Financial Times calling for the establishment of a US-led international AI safety forum. The forum would be modeled on the structure and operations of the International Atomic Energy Agency (IAEA), aiming to create global standards for AI model safety testing and governance among democratic nations. This marks a significant step in bringing AI regulation back onto the international policy agenda following the G7 summit.
Proposal Background and Timing
The proposal was released after the G7 summit held on June 17, 2025. According to public reports, the Trump administration held closed-door meetings with AI executives during the summit to discuss US-led global AI regulation plans. Altman's proposal is seen as an industry response to the government's policy direction, pushing for a binding multilateral safety framework.
Forum Positioning: Differences from the IAEA
Notably, unlike the IAEA which primarily addresses confirmed nuclear threats, the proposed AI safety forum would focus on forward-looking risk governance—specifically, pre-deployment safety testing and evaluation of AI model capabilities. Altman emphasized in his proposal that AI technology evolves far more rapidly than traditional nuclear technology, necessitating a preventive multilateral cooperation mechanism rather than a reactive one after incidents occur. The forum would set unified testing standards, certification processes, and transparency requirements, prioritizing implementation among democratic nations.
Potential Implications for the Crypto Industry
Although the proposal remains at a conceptual stage, it has already sparked widespread discussion in the crypto industry and tech policy circles. The convergence of AI and blockchain—such as decentralized AI inference and on-chain model verification—has been a hot topic in recent years. The establishment of a global AI safety regulatory framework could indirectly influence the development paths of decentralized AI projects. For example, if the forum requires AI models to undergo specific safety certifications, models deployed on-chain may face compliance challenges. Additionally, crypto industry analysts believe that a US-led regulatory model could accelerate competition and cooperation among nations in AI governance, thereby affecting market expectations for related tokens and protocols.
As of now, OpenAI has not released further details on the proposal. The responses from the Trump administration and G7 member states remain to be seen. However, the proposal marks a critical turning point where AI safety governance moves from industry self-regulation toward intergovernmental cooperation. Its subsequent progress warrants close attention from the global crypto community.