Two thirds of Britons think artificial intelligence could bring about the end of human civilisation, according to new YouGov polling, with a quarter saying they expect it to happen.
The figures land after a week of warnings from inside the industry itself. Jacob Coxon, a former Anthropic employee, quit with a resignation thread on X accusing both OpenAI and Anthropic of “gambling with our lives”. A colleague, Evan Hubinger, publicly agreed and put the odds of an AI-driven apocalypse at greater than 10%. Anthropic’s chief executive Dario Amodei then published an essay of his own, backed by Sam Altman of OpenAI, Demis Hassabis of Google DeepMind and Elon Musk of xAI.
The public were already worried. A third of Britons (34%) now name robots and AI as one of the three most likely causes of human extinction, more than double the 15% who said so in August last year. Pessimism about the everyday effects has climbed faster still. Some 52% expect AI’s overall impact to be negative, up from 34% in 2024. Just 19% are optimistic.
Nobody trusts the people building or policing AI
The distrust does not stop at the industry. Eight in ten Britons (79%) have little or no confidence in AI companies to develop the technology responsibly. Eight in ten (80%) say exactly the same about the ability of this government, or any future one, to regulate them.
Almost three quarters (73%) say AI firms face too little regulation, including 51% who say “far too little”. Seven per cent think the current rules are about right. One per cent think there is too much red tape.
The public want a slow down
Given six options on the pace of development, Britons picked the slow lane. Six in ten (61%) back slowing AI down while letting it continue, against 24% opposed. Half (51%) would support a temporary pause.
Four in ten (40%) would stop new development permanently and let people use only what has been built so far. Nearly three in ten (29%) would ban the technology outright.
Carrying on as we are attracts 21% support and 62% opposition. Speeding up is rejected by 75%.
Not a party-political issue
Voters for all five of the largest parties at the 2024 election take a gloomy view of AI, want tougher regulation, distrust the companies and ministers alike, and want the pace slowed. They disagree about almost everything else. They do not disagree about this.
No party has managed to speak for them. Asked which would handle AI best, none scored above 9%. Thirty-nine per cent said they did not know. Twenty-seven per cent said no party would be capable of managing it at all.
Nor is this squeamishness from people who have never touched the stuff. Daily users are more cheerful about the technology, with 44% optimistic. They are still not defending the industry. Three quarters of them (75%) say it faces too little regulation and 70% lack confidence in AI companies to behave responsibly. Roughly two thirds want development slowed.
The bold promises are not landing
Amodei’s essay offered a prospectus: “I believe that AI could cure most major diseases in the next 5-10 years, greatly accelerate economic growth rates, create a world of abundance and empowerment”.
Britons are not buying it. Half (51%) accept AI could cure most major diseases, but only 12-13% think it likely to happen. On the promise of a better democracy, 51% dispute the idea in the first place. Five per cent think it likely.
One in a hundred of us believes AI companies are tied up in too much red tape. Nobody in Westminster has yet proved the other ninety-nine wrong.