Dario Amodei, CEO of Anthropic, has proposed a three-stage plan to moderate the development of artificial intelligence. The initiative includes granting independent experts access to internal safety procedures, coordinating safety standards among democratic nations, and establishing international agreements between states with different political regimes.
The proposal received support from Sam Altman of OpenAI and Elon Musk of xAI. Altman confirmed that OpenAI is considering similar measures, including external expert oversight. Additionally, OpenAI has decided to postpone its planned 2026 IPO to prioritize safety improvements.
Anthropic's public relations representative, Sarah Heck, urged the US government to regulate the industry. The company advocates for a ban on the export of advanced chips to China and the implementation of mandatory safety testing for new AI models.
The discussion on slowing down AI development follows criticism from former employees of Anthropic and OpenAI regarding safety protocols. Amodei noted that while the potential for AI to improve human life remains, the industry must prioritize safety to mitigate long-term risks.
