- AI Pioneer Geoffrey Hint warns This artificial intelligence can one day be overcome and controlled by humanity. He says that a profit-based AI-weapon competition can accelerate the danger.
Geoffrey Hinton, one of the “AI godfathers,” warns that the technology he helped to create could take control of people.
In an interview With CBS On Saturday morning, Hinton said that AI had developed even faster than he had originally expected, referring to the acceleration of AI agents as a particularly scary development.
He also predicted that artificial general intelligence, theoretical AI system, which has human cognitive abilities, could arrive in 10 years.
Hinton, one of the creators of the main neural network technologies, has long warned that AI may pose an existential threat to people.
In 2023, he left Google after he was deeply concerned about the risks that artificial intelligence (AI) could be portrayed on humanity. At that time, he said he wanted to talk about these dangers in public without being limited to a larger technology company.
He has previously warned that AI systems may one day become smarter than people and act in a way that we cannot control – potentially threat to humanity itself.
In an interview aired at the end of last week, Hinton doubled these concerns.
“Given the possibility that these things will become much smarter than us, and then just take control of us, just take over, the likelihood that it is likely to be more than 1% and probably less than 99%. All experts can accept it,” he said.
“I am in a regrettable situation when I agree with Elon Musk, that is, it is a kind of 10% to 20% probability that these things will take over,” he said, adding that it was still just “wild guessing.”
“The best way to understand this emotionally is that we are like someone who has this really cute tiger store,” Hinton explained. “If you can't be very sure that it doesn't want to kill you when you grow up, you should worry.”
He emphasized that as the world continues to approach AI with a profit-based mindset, AI was more likely to take over or bad actors who chose technology for dangerous tools such as mass supervision.
“When you look at what big companies are doing right now, lobbying to get less AI regulation. Hardly as it is, they are hardly, but they want less,” Hinton said. “We have to put pressure on governments to do something serious about it.”
Hinton is not the only AI godfather to talk about the technology he helped to create. Yoshua Bengio, who won the Turing Award in 2018, has also warned that AI could overcome people and potentially take over.
Bengio has gone a little further from AI's threats, saying he feels he feels to invent guilt Deep learning, because it can now be abused in a dangerous way.
AI influence on the world
Despite his concerns, Hinton said that there is also reason to be optimistic about the influence of AI on the world, pointing to health, medication and education as areas where AI could improve human experience.
AI has already shown signs of reaching A The human level of diagnostic skillsBut it is below other regions.
In education, the rise of AI tutors could help students personalize education and unload some teachers' workload. Openai is already as a partner with several leading Universities give students and employees access to funding and “top-level” AI tools.
Hinton said that the rise of personalized AI tutors would be “bad news for universities, but good news for people”.
He also changed his position on AI's resettlement of the job, saying that he is now worried that people could lose their jobs due to technology and warned that it could exacerbate the difference in wealth.
This story was originally reflected on Fortune.com
