Coordinated hunger strikes have called on Big Tech to stop developing ‘vastly superhuman’ artificial intelligence which could destroy us.
Dad-of-two Guido Reichstadter is now in his 22nd day without food outside Anthropic in San Fransisco, saying ‘the world’s AI companies are driving us headlong into a minefield’.
After seeing what he was doing, Michaël Trazzi, 29, was inspired to do the same outside the London offices of Google’s DeepMind research lab. Another protester, Denys Sheremet, then travelled from Amsterdam to join him two days later.
Former AI researcher Mr Trazzi, 29, told Metro they ate zero calories and waited outside the officers from 9am until 7pm, coinciding with the working day.
He said that after episodes where he nearly fainted, doctors advised him to stop after a week after tests showed dangerously low blood sugar, putting him at risk of risk of ‘seizure, brain damage, or death if I continue’.
Denys kept going until Monday, lasting 16 days with only water and electrolytes.
Their concern is that the race to develop frontier AI puts humans at existential risk, from superintelligence in the longterm, and from terrorists getting new tools in the shortterm, like the ability to engineer super-viruses and bioweapons.
Companies like OpenAI and Google are working towards building Artificial General Intelligence (AGI), which would could do everything a human can do, and much more.
But we don’t know how something much smarter than us would behave, or how we could control it, and humans could be collateral damage in the goals it pursues.
As Denys puts it: ‘When we build a house, we don’t go around and ask all the ants for permission. We just build a house, and the ants might get unlucky in the process.’
Michael’s interest in DeepMind is personal, as he saw the potential of his models early on when AlphaGo beat one of his Go teachers in 2016: its mastery of the complex game was a milestone in AI development. as it required strategic thinking.
Latest London news
- Three children injured after ‘war zone’ fire rips through flats
- Gatwick Airport’s second runway expansion given go-ahead
- Lime bike gangs are targeting luxury cars across London
To get the latest news from the capital visit Metro’s London news hub.
After a stint at Oxford’s Future of Humanity Institute, he now works as a film maker raising awareness of the risks of the trillion-dollar AI industry.
He says he isn’t singling out DeepMind as the worst offender; in fact, he views the company as historically having deep concern for safe development of the technology.
But the current race for AI dominance is making it hard for any company to stop and evaluate, he said, claiming their public statements on safety are at odds with reality.
In a letter to DeepMind CEO Sir Demis Hassabis, he urges him ‘to publicly state that you will halt the development of frontier AI models if all the other major AI companies do the same’.
Letter to DeepMind founder and CEO Demis Hassabis
Dear Demis Hassabis,
I’m on Day 5 of a hunger strike outside Google DeepMind’s headquarters in London, asking you to publicly state that you will halt the development of frontier AI models if all the other major AI companies do the same.
I believe that when you started DeepMind you were truly committed to building safe Artificial General Intelligence, but there is now a contradiction between your public statements on AI safety and your continued race towards superintelligence through the release of ever more powerful AI models. This race may end in self-improving AI that is beyond our ability to control, according to the three most cited AI researchers in the world.
I understand that there are strong financial and competitive incentives for DeepMind to continue pushing the frontier. And I’m also aware that there are many potential applications of AIs that would be beneficial to humanity, such as medical AI that could cure diseases.
Which is why I’m asking you to take a first step today towards coordinating a future halt on the development of superintelligence, by publicly stating that DeepMind would agree to halt the development of frontier AI models if all the other major AI companies in the West and China were to do the same. Once all major companies have agreed to a pause, governments could organise an international agreement to enforce it.
Yours sincerely,
Michaël Trazzi
The protesters are not just talking about some fringe conspiracy or anticipating doom like the Rapture: AI executives have repeatedly acknowledged the risk of human ‘extinction’ too.
In 2023, Demis Hassabis co-signed a statement alongside Anthrophic CEO Dario Amodei, OpenAI boss Sam Altman, and dozens of other experts warning that ‘mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.’
Lately however, the push for supremacy only seems to be hotting up: there was little talk of slowing capabilities when the UK government announced billions of pounds in investment in the tech last week.
Two days ago, DeepMind published its latest Frontier Safety Framework (FSF), calling it their ‘most comprehensive approach yet to identifying and mitigating severe risks from advanced AI models’, saying they were ‘committed to responsibly developing our technologies and taking an evidence-based approach to staying ahead of emerging risks’.
Andrea Miotti, of ControlAI, told Metro: ‘We’re seeing more and more people undertake protests of this kind, as the public increasingly demands accountability from the leading AI developers. Millions around the world are learning about the extinction threat posed by superintelligence, but most people don’t know how to make a difference.
‘The one thing everyone can do is contact their lawmakers and demand action to prevent the development of superintelligence.’
Get in touch with our news team by emailing us at webnews@metro.co.uk.
For more stories like this, check our news page.
