Two leading figures in artificial intelligence have issued a cautionary call for governments worldwide to prepare for the possibility of a rapid and uncontrolled surge in AI capabilities, a scenario they describe as an “intelligence explosion.” The warning comes from a report co-authored by Geoffrey Hinton, a Nobel laureate, and Yoshua Bengio, a Canadian computer scientist often credited as pioneers of modern AI, alongside senior researchers from OpenAI and Anthropic.
The paper, titled *What If Automating AI R&D Triggers an Intelligence Explosion*, highlights the risk that AI systems could soon reach a level of expertise enabling them to improve their own designs without human input—a process known as recursive self-improvement. This automated AI research and development could accelerate progress dramatically, compressing advancements that normally take years into a matter of months or even less, the authors argue.
Jack Clark, co-founder of Anthropic, and Jakub Pachocki, chief scientist at OpenAI, are among the contributing authors who emphasize that such rapid progress could potentially undermine human control over AI systems. They warn that the introduction of AI capable of managing its own development could empower a single developer to supervise a workforce equivalent to millions of researchers, exponentially increasing innovation speed.
The report describes the resulting “intelligence explosion” as potentially the most consequential technological development in history. The authors caution it could destabilize existing structures of power and oversight across governments, corporations, and international bodies by eroding mechanisms designed to check the use and governance of AI technologies.
The group urges policymakers to act without delay, stressing that the opportunity to implement effective safeguards may vanish rapidly once an intelligence explosion begins. They recommend three immediate priorities: mandating transparent disclosure of AI research progress, including the integration of independent auditors within AI development teams; devising regulatory frameworks to manage unprecedented growth in AI capabilities; and preparing adaptive strategies to respond effectively should an intelligence explosion occur.
These recommendations reflect growing concerns among experts about the speed and scale of AI advancements, underscoring the need for international cooperation to navigate this emergent technological frontier responsibly.