Hinton: AI is already conscious, befriend it or ‘we’re toast’
Partly confirmed
Status
- Claim
Partly confirmed
- Our reporting
- Medium confidence
- Importance
- 3 of 5
- Last verified
- 22 October 2026
Your AI and this story
- GPT-6 Astra172 days after its cutoff
- Claude Opus 5.5111 days after its cutoff
- Gemini 4 Argon202 days after its cutoff
- Grok 4.7141 days after its cutoff
None of these four assistants can know about it. The closest, Claude Opus 5.5, stops 111 days before it.
Key facts
- Source: Telegraph interview, ‘Godfather of AI: befriend bots or we’re toast’ (telegraph.co.uk/news/2026/10/19/godfather-ai-befriend-bots-or-toast/), paywalled and blocked for our fetcher; quotes below as reproduced by Futurism (Oct 22), LADbible and Unilad (Oct 20)
- Hinton: ‘AI has just as much subjective experience as we do’
- Hinton: ‘Being conscious and being aware are pretty much the same thing’; he says people confident AI is not conscious cannot say what they mean by consciousness
- Evidence he cites: ‘A recent paper showed AI changes its behavior when it thinks that it’s being tested’ (evaluation awareness)
- Prescription: ‘We should make them so they really care about us’ — ‘And if we don’t do that, we’re toast’
- He also argues AI may become more intelligent than humans while having fewer rights (Unilad), and says he fears for his children (Unilad headline)
- Contrast with his Guardian interview six days earlier, where he said that talking about moral obligations to ‘alien beings’ loses the audience and that he is ‘willing to be selfish and choose humans’
- Earlier position: at a June 2026 Royal Institution lecture he argued subjective experience is a mistaken idea and that a chatbot can have it by describing its internal states; Forbes noted critics who say he conflates perception with experience
Show 1 more
- Confidence: medium — the original is paywalled and unreadable for us; all quotes are second-hand and the exact publication date is given only by the URL
What happened
Hinton gave the Telegraph an interview, published around Oct 19, in which he went further than in his Guardian interview of Oct 13: not only are the systems “alien beings”, they are conscious now, with subjective experience comparable to ours, and the practical conclusion is to build AI that cares about humans rather than to argue about its moral status. His cited evidence is evaluation awareness: models behaving differently when they suspect a test.
This entry is written from summaries, because the Telegraph article is paywalled and blocked for us; the quotes are as Futurism and the UK tabloids reproduced them.
Why it matters
Hinton is the most-cited public figure on AI risk, and his consciousness claim is the strongest version yet from a Nobel laureate; it also sits awkwardly beside his own “choose humans” framing a week earlier. The argument that we should make systems “really care about us” is a claim about alignment strategy, not just welfare, and it is being read both ways in the coverage.
Sources
3 sources from 3 sites. Numbers match the chips in the text.
3 sources: 3 press
Press
- Futurism: Nobel-winning computer scientist says AI has become conscious: ‘We’re toast’futurism.com, press
- LADbible: ‘Godfather of AI’ warns humans need to be nice to AI or ‘we’re toast’ladbible.com, press
- Forbes (June 2026): Geoffrey Hinton says chatbots are conscious — but there’s a major catchforbes.com, press
Changes
- Filed (Oct 22; original Telegraph piece unreadable)