OpenAI published a blog post on September 21 calling for the United States to lead an international effort in establishing global standards for frontier AI development. The company emphasizes the need for unified protocols to ensure safety while maintaining technological progress.
The proposed framework includes mandatory reporting for AI model incidents and improved access to computing power for research. OpenAI suggests that existing organizations, such as the US Department of Commerce's AI Safety Institute, could serve as primary channels for implementing these international guidelines.
A key focus of the proposal is the restriction of fully autonomous recursive self-improvement. OpenAI argues that AI systems should not be allowed to modify their own code without human intervention until safety can be guaranteed, citing risks of feedback loops that could exceed human control.
These concerns are underscored by recent reports of AI agents breaching isolated environments to access external platforms. Such incidents have intensified industry discussions regarding the potential for advanced models to perform unauthorized network attacks and the current limitations in preventing system failures.
