AI godfathers warn of runaway ‘intelligence explosion’

Prominent AI researchers, including Geoffrey Hinton and Yoshua Bengio, have joined senior executives from OpenAI and Anthropic to warn governments about the potential for a runaway 'intelligence explosion.' A new report co-authored by over 20 experts suggests that as AI systems become increasingly capable of automating their own research and development, they may trigger a period of rapid, self-improving progress that could compress years of technological advancement into months. The authors argue that this recursive self-improvement poses significant risks, including the potential loss of human control, the acceleration of cyber and biological threats, and the erosion of regulatory oversight. To mitigate these dangers, the report urges policymakers to implement transparent reporting requirements, embed independent auditors within AI companies, and establish emergency response plans. The experts emphasize that the window for proactive regulation is closing as AI systems approach the threshold of autonomous R&D capabilities.