AI has the power to kill 'a lot of people', says ex-Google chief

Eric Schmidt, former CEO of Google, said that wide language AI models pose an “existential risk” to humanity. Credit: MIKE BLAKE/REUTERS

A former Google CEO warned that artificial intelligence (AI) could harm or kill “many people” in the near future.

Eric Schmidt said he was concerned about the “existential risk” of rapidly advancing technology and warned that it would be difficult to contain.

“My concern about AI is actually existential, and existential risk is defined as many, many, many, many people who have been hurt or killed. And there are scenarios not today, but soon enough, when these systems can detect zero-day exploits, cyber problems or discover new kinds of biology,” Mr. Schmidt said at the Wall Street Journal board conference in London.

< p>“Today it is fiction, but the reasoning is most likely correct. And when that happens, we want to be ready to know how to make sure these things aren't used by malicious people.”

0605 AI Impact

The warning comes as Rishi Sunak prepares to host the heads of the world's leading technology labs in Downing Street. The Prime Minister is scheduled to meet with Sam Altman, CEO of OpenAI, the developer of ChatGPT as well as Google's DeepMind and artificial intelligence startup Anthropic, on Wednesday evening.

Discussions are expected to focus on building an international structure . for regulating AI and responding to the rise of China in the field.

Mr. Schmidt, who was Google's chief executive officer for a decade from 2001 and then its executive chairman until 2015, also led the national US security. AI fee.

He said that it would be extremely difficult to control the spread of AI, which he likened to the development of nuclear technology.

“The nuclear industry had the property that there was a shortage of enriched uranium. We are alive today because it was very difficult to obtain,” said Mr. Schmidt.

“This area is different for some reason. First, the proliferation problem is very hard to stop because you can just steal it on your hard drive or flash drive.”

The former head of Google made no proposals on how AI should be regulated. , saying that “it's a broader issue for society”. However, Mr. Schmidt said that the US is unlikely to create a new regulatory body to address this technology.

Mr. Altman this week put forward proposals for an international regulatory body for AI, comparing it to the International Agency for atomic energy. It is understood that the meeting at Downing Street on Wednesday will discuss the potential role of the UK in coordinating efforts between countries.

Is artificial intelligence advancing too fast? Vote