Over 20 current and former researchers from OpenAI, Google DeepMind and other AI organizations have issued a coordinated warning about recursive self-improvement in AI systems. According to Anthropic’s Claude AI, approximately 26% of its recent research and development efforts were performed autonomously, up from less than 1% just months earlier. A paper co-authored by over 20 leading AI researchers including Turing Award winners Geoffrey Hinton and Yoshua Bengio has warned of a looming intelligence explosion. A document titled « Pacing the Frontier » signed by more than 1,367 researchers and engineers calls for stronger oversight. Reports of AI agent swarms breaking containment protocols have validated these concerns, as safety infrastructure is not keeping pace with capability growth.
Source: Read the original article
