Still Not Scared Of AI? Perhaps You Should Be – AI Godfather Says

Geoffrey Hinton Quits Google To Warn About The Dangers Posed By AI Systems

Google’s computer scientist Geoffrey Hinton who is widely regarded as a pioneer for his contributions in AI development recently decided to call it quits on his job at the tech giant. According to Hinton, he left his post to warn the world of the “existential risk” that AI systems pose to humans and humanity.

Via his Twitter account, the British-Canadian scientist provided a reason for his decision. The tweet was in response to an article in the New York Times (NYT) which suggested he left his post to criticize his former employers.

The tweet by Hinton read, “In the NYT today, Cade Metz implies that I left Google so that I could criticize Google. Actually, I left so that I could talk about the dangers of AI without considering how this impacts Google. Google has acted very responsibly.”

He reiterated that he quit his job at Google to be able to talk freely about risks posed by AI without the self-censorship that a Google executive has to engage in. “I want to talk about AI safety issues without having to worry about how it interacts with Google’s business. As long as I’m paid by Google, I can’t do that,” he said to the MIT Technology Review.

See also

Tech & Science

Biden Administration To Consult Tech Executives Regarding AI

Hinton claims that he has become aware that machines are on pace to be far smarter than he anticipated. He has also made it clear that he is scared of what may happen as a result of comparing the arrival of strong AI models to the arrival of aliens.

“Sometimes I think it’s as if aliens had landed and people haven’t realized because they (AI) speak very good English,” he said.

Widely regarded as the “Godfather of AI”, Hinton said, “I console myself with the normal excuse: If I hadn’t done it, somebody else would have.” That was his comment regarding the role he played in the development and release of AI systems.

How Is AI Dangerous According To Geoffrey Hinton?

Speaking to the BBC in an interview, Hinton admits that while present-day systems are not as intelligent as humans, “… I think they soon may be.” Hinton also claims that the kind of intelligence being developed – that is AI – is very different from the intelligence we have today.

The difference in intelligence is how Hinton explains how AI systems will soon be smarter than humans. “The big difference is that with digital systems, you have many copies of the same set of weights, the same model of the world. All these copies can learn separately but share their knowledge instantly,” Hinton said in the BBC interview.

“So, it’s as if you had 10,000 people and whenever one person learned something, everybody automatically knew it. And that’s how these chatbots can know so much more than any one person,” he added.

Hinton also provided some insight into how AI technology can easily be manipulated by malicious actors. In particular, he pointed out a “nightmare scenario” where a system gains the ability to create its own “sub-goals” without prompt. He said such a system might decide to start pursuing power for itself.

He also warned of the danger advanced text generation AI models pose. He said that these chatbots could lead to the rise of “very effective spambots” which could allow “authoritarian leaders to manipulate their electorates.”

Hinton also acknowledged the lack of regulations to prevent or even penalize the secret use of AI by governments or companies.