Several prominent artificial intelligence researchers have publicly supported a warning from an Anthropic researcher that advanced AI systems could cause catastrophic, irreversible loss of control with severe human consequences.
Paul Christiano — who co-authored an influential 2016 paper on AI safety with Dario Amodei, a future Anthropic founder — previously believed AI would likely develop relatively gradually and be made safe. He now says he sees a “meaningful risk” of “catastrophic and irreversible loss of control in the very near term.” Christiano added the stark warning that “most people could die,” and announced he is joining OpenAI’s nonprofit safety team to work on reducing these risks.
Geoffrey Hinton, often referred to as one of the three “Godfathers of AI,” told the BBC that a 10% risk of human extinction “seems a not unreasonable estimate.” That comment echoes a warning made by an Anthropic senior AI alignment scientist the previous day.
Why this matters
These statements carry weight because of the authors’ professional standing and can influence public debate, regulatory attention, and research priorities in AI safety. The central concern is that a rapid loss of control over powerful AI systems could produce consequences that are both severe and difficult or impossible to reverse.
Concrete actions
As of the announcements, Paul Christiano is joining OpenAI’s nonprofit safety team to help mitigate identified risks. Beyond that, the likely outcomes include intensified discussion about governance measures, potential regulatory initiatives, industry self-regulation, and increased funding for AI safety research—depending on how policymakers, companies, and the research community respond.
Summary
The endorsements from senior AI researchers such as Paul Christiano and Geoffrey Hinton underscore that a growing number of experts treat the safety risks from advanced AI as real and urgent. Their remarks are likely to accelerate debate and action on how to govern and control powerful AI systems.
