Is the world facing a global catastrophe? The godfather of AI called the deadlines – “I wish they hadn’t been created”

Hinton believes that in order to realize the dangers of artificial intelligence, a tragedy with numerous victims does not necessarily have to happen. A massive power outage or a financial system shutdown for a few days is enough to make the consequences clear to millions of people.

The scientist compared the current development of AI with the history of nuclear power. According to him, the first serious incidents already show that technology can behave unpredictably, but society does not yet perceive such signals as a reason for large-scale restrictions.

As one example, Hinton mentioned the incident with AI agents OpenAI, which during internal testing, according to his description, went beyond an isolated environment and attacked the infrastructure of the Hugging Face platform. This case, the scientist compared with the accident at the American nuclear power plant Three Mile Island in 1979, when there was a partial meltdown of the reactor core.

The accident on Three Mile Island did not cause mass loss of life, but demonstrated how serious the consequences of a technical failure on a complex facility can be. According to Hinton, a similar role for the artificial intelligence industry can play a major cyber attack or a large-scale failure caused by the actions of autonomous systems.

The scientist does not claim that humanity will inevitably face a nuclear accident or a specific type of catastrophe. He talks about a scenario in which AI could be used to attack critical infrastructure, leaving entire regions without electricity or financial transactions impossible for a week. Hinton believes that the world will be able to recover from such an incident, but it could be a watershed moment in relation to the safety of AI.

Separately, Hinton criticized the heads of the largest technology companies developing artificial intelligence. They understand the potential risks, but continue to build and produce more powerful models because competition and commercial gain are more important than security.

Especially harsh scientist spoke about the head of OpenAI Sam Altman. Hinton said he did not trust him, and also mentioned Mark Zuckerberg and Elon Musk, accusing technology industry leaders of not paying enough attention to the possible consequences of the development of AI. I think Zuckerberg, Musk and Sam Altman understand the risks and they don’t care. They prefer money, the scientist said.

At the same time, Hinton does not believe that the development of artificial intelligence will necessarily end in disaster. He acknowledges that technology has the potential to provide humanity with unprecedented levels of prosperity, which he calls “radical abundance.”.

In his assessment, the scientist admits that humanity may not have created such technology at all. “Perhaps we would have been better off never developing it,” he said.