AI 'Godfathers' Hinton and Bengio Warn Governments of Intelligence Explosion

Two of the three men widely credited as the "godfathers" of modern AI have joined senior executives from OpenAI and Anthropic in warning governments to prepare for an AI "intelligence explosion" — a prospect they describe as potentially the most consequential technological development in history.

The warning comes in a report co-authored by Nobel laureate Geoffrey Hinton and Canadian computer scientist Yoshua Bengio, according to The Guardian. The report urges politicians to act now, before runaway progress in the technology outstrips the ability of governments to respond.

Hinton and Bengio earned their "godfather" status through decades of foundational work on the neural network techniques that underpin today's large language models. Hinton received the Nobel Prize in Physics in 2024 for that work. Both have since become two of the most prominent internal critics of the field they helped build, and their intervention carries a weight that few other voices in the AI debate can match.

The chief scientist of OpenAI is also among the report's authors, The Guardian reports. That detail matters. It places a senior figure from the industry's leading lab alongside its most credentialed external critics, on the same page, warning of the same outcome. An Anthropic executive is listed among the authors as well.

What the report says

The core claim is direct: governments must prepare for the possibility of an intelligence explosion — a scenario in which AI systems become capable of improving themselves, accelerating progress beyond human control or comprehension. The authors frame this not as a distant theoretical concern but as a live policy problem that demands preparation before the runaway dynamic begins, not after.

The report urges politicians to act now. The authors call the intelligence explosion potentially "the most consequential technological development in history."

Why the messengers matter

The composition of the author group is itself part of the story. Hinton left Google in 2023 and has repeatedly raised his estimates of the risks posed by advanced AI, including the possibility of the technology contributing to human extinction within the coming decades. Bengio has become one of the most active scientific voices calling for government regulation of frontier AI development.

Their co-authors now add industry's perspective from inside the labs themselves. When the people building frontier systems at OpenAI and Anthropic sign a document alongside the field's most famous external skeptics, the usual industry-versus-critics framing collapses. The disagreement is no longer about whether extreme outcomes deserve serious planning. It is about what governments should actually do, and how fast.

The policy stakes

The report lands at a moment when governments are actively debating how — and whether — to regulate frontier AI. The Guardian's coverage notes that the political environment is contested. In September 2026, US Vice President JD Vance dismissed AI regulation in remarks comparing the technology to "building Frankenstein," while separate reporting the same month detailed government action on AI regulation urged by Bengio himself.

That is the context the report's authors are speaking into. The dividing line in Western capitals is no longer between AI skeptics and AI enthusiasts. It is between those who believe market forces and existing law are sufficient, and those — now including senior lab staff — who argue that governments need preparation frameworks in place before an intelligence explosion becomes a live possibility rather than a hypothesis.

The report's central policy argument follows from its central technical claim. If progress can become self-accelerating, then waiting for clear evidence of the transition before building response capacity means waiting too long. Preparation, in the authors' framing, has to happen while the outcome remains uncertain — which is precisely what makes it politically difficult.

The significance of the timing

The intervention follows a period of escalating public warnings from Hinton. In December 2024, he raised his estimate of the odds of AI wiping out humanity over the following 30 years. His position has grown steadily more alarmed since he left Google, and this report represents a formalization of that warning into a direct appeal to political leaders.

Bengio's involvement is equally consistent with his trajectory. He has spent the period since the ChatGPT boom calling for binding government oversight of frontier model development, positioning himself between an industry resistant to hard rules and critics who argue the entire enterprise should be paused.

What is new is the alliance. A document that pairs Nobel-laureate academics with the chief scientist of OpenAI and a senior Anthropic executive sends a signal that the probability assessments behind extreme-outcome scenarios are shared, at least in part, by people with direct visibility into the most advanced systems in existence.

Why this matters

For policymakers, the report reframes the AI governance debate around a specific question: what concrete preparation — monitoring, evaluation regimes, emergency response capacity, international coordination — should exist today to handle a scenario in which AI capabilities begin improving themselves? The authors' answer is that the window for building that capacity is open now and may not stay open.

For the industry, the participation of OpenAI and Anthropic leadership in the warning creates pressure to demonstrate that internal safety practices are commensurate with the risks their own scientists are describing.

For the public, the report is the clearest statement yet that the founders of the deep learning era and the builders of its most advanced products agree on one thing: the trajectory of AI progress is not guaranteed to remain under human control, and governments are not yet prepared for the scenario in which it does not.

The report's authors have made their recommendation explicit. Whether political leaders act on it before, as they warn, runaway progress makes preparation impossible is now the open question that will define the next phase of the AI policy debate.