On September 29, 2026, leaders from major artificial intelligence companies, including Anthropic, OpenAI, Meta, and Microsoft, called for increased oversight of self-improving systems. They expressed concerns about the potential risk of an "intelligence explosion," which could lead to rapid technological advancements that outpace society's ability to adapt.
Risks of Research Automation
In their statement, over 20 leaders and researchers emphasized that involving artificial intelligence in the creation of new models could pose safety threats. They believe that the accelerated development of technology may result in diminished human oversight, creating risks for society.
Proposals for Enhanced Control
The authors of the call suggest several measures to enhance safety in the field of artificial intelligence:
- Assess the extent of research automation in leading laboratories.
- Strengthen security and oversight of data centers involved in AI development.
- Prepare emergency response plans for potential threats.
Notable Signatories of the Statement
Among the signatories of this statement are well-known figures in the field of artificial intelligence, such as Jack Clark (co-founder of Anthropic), Jakub Pachocki (chief scientist at OpenAI), Geoffrey Hinton, and Yoshua Bengio. Their stance underscores the importance of caution in the development of technologies that could change the world.