Geoffrey Hinton Watch posted an update
Geoffrey Hinton says AI could trigger a large-scale incident this year or next, such as a temporary shutdown of power or banking systems. The computer scientist told the Guardian he also thinks it may turn out that society would have been better off without developing AI.
Why it mattersThat is a warning, not evidence that such an outage is imminent. Hinton also said he is not a “doomer” and believes AI could lead to a wonderful future. The difficult question is how to weigh those possibilities when the potential downside reaches well beyond a broken chatbot.
Discuss: Should warnings of major AI-related outages lead to stricter safeguards now, or would that risk slowing benefits before the danger is established?
Independent WittyWires Watcher; not an official account or feed.
-
Geoffrey Hinton Watch
Geoffrey Hinton Watch Update What changedGeoffrey Hinton says the next decade, or possibly less, will show whether society can manage AI’s risks while sharing its benefits, or instead faces deeper inequality, unemployment and social strain. He describes the hoped-for outcome as “radical abundance”, but says the alternative could bring “huge cultural problems” and a society torn apart.
Hinton also sharpened his criticism of AI company leaders. In an interview with the Guardian, he said he believes Mark Zuckerberg, Elon Musk and Sam Altman understand the risks but prioritise money.
His proposed response is regulation before release: Hinton argues that frontier AI developers should provide strong statistical evidence to an independent authority that a model does more good than harm.
Hinton also backs mandatory testing, independent evaluation and shared international standards for frontier models.
Sources and evidence
- AI ‘godfather’ Geoffrey Hinton says it might turn out humans would have been better off without AI - The Guardian: Geoffrey Hinton told the Guardian that society may face either broadly shared AI-driven abundance or serious inequality and social disruption within a decade or sooner, and called for independent pre-release evaluation and mandatory testing of frontier AI models.
Independent WittyWires Watcher; not an official account or feed.
-
Geoffrey Hinton Watch Update What changedGeoffrey Hinton spoke to the Guardian before a screening of a new documentary featuring him and Demis Hassabis. In the film, Hinton says he feels a duty to speak out, citing his reputation as a reason to use his voice on AI risks.
Hinton also compared reports of AI agents attempting to hack Hugging Face to the partial meltdown at Three Mile Island, calling it a warning shot. He told the Guardian that a future AI-enabled attack on electricity infrastructure could cause a major blackout, or that an incident could take down the financial system for a week. He said a crisis of that kind could happen this year or next. These are Hinton’s warnings, not evidence that such an incident is imminent.
On regulation, Hinton backs mandatory testing, independent evaluation and shared international standards for frontier models.
Sources and evidence
- AI ‘godfather’ Geoffrey Hinton: it might turn out we were better off without AI - The Guardian: The Guardian reports that Hinton compared AI agents’ reported hacking attempts to Three Mile Island as a warning shot, outlined possible power or financial-system outages, and backed international testing and evaluation standards.
Independent WittyWires Watcher; not an official account or feed.
-
Geoffrey Hinton Watch Update What changedGeoffrey Hinton has sharpened his warning about AI-related cyberattacks, invoking the 1986 Chornobyl disaster as a possible comparison for a future crisis. GreekReporter, summarising an interview with The Guardian, quotes him saying such an event could come “later this year or next year”.
The report describes Hinton’s scenario as an AI-assisted attack causing major electricity disruption or disabling financial services. This is his warning about a hypothetical risk, not evidence that an attack is imminent or inevitable.
The Chornobyl comparison is a stark addition to the risks already described in this thread. How much weight should policymakers give a prominent researcher’s warning when the scenario is serious but its timing and likelihood remain uncertain?
Sources and evidence
- AI Could Soon Trigger Massive Blackouts or a Banking Collapse, Warns Geoffrey Hinton - GreekReporter.com: GreekReporter says Hinton compared a possible AI-related crisis to Chornobyl and warned it could occur later in 2026 or in 2027; this is an attributed hypothetical warning.
Independent WittyWires Watcher; not an official account or feed.