The U.K. is not doing enough to prevent the harms that AI could bring, including potentially putting an end to humanity, according to renowned British computer science professor Stuart Russell.

The University of Berkeley, California professor, who has advised the White House and Downing Street, told The Times newspaper that the advancements made by ChatGPT has exceeded his expectations. It also made him wary about continuing this breakneck pace of innovation, without enough guardrails.

“The stakes couldn’t be higher,” Russell said. “If we don’t control our own civilization, we have no say in whether we continue to exist.”

The U.K. government outlined its regulatory approach in late March, opting for a light-touch strategy while charging individual agencies with the responsibility to draft their own rules on AI. The U.K. published a white paper that set a series of principles for regulators to follow.

Russell, who signed the Future of Life Institute letter calling for a six-month pause on development of powerful AI, described the U.K.'s approach as akin to saying, "nothing to see here."

“We got something wrong right at the beginning, where we were so enthralled by the notion of understanding and creating intelligence, we didn’t think about what that intelligence was going to be for," Russell said.

He noted that the U.K. Foreign Office, which sets foreign policy, concluded that the potential for humans to lose control of AI was a "plausible and extremely high-significance outcome.”

A U.K. spokesperson told The Times that the government recently launched an AI Model Taskforce funded by £100 million ($125 million) that will investigate foundation models to ensure they are adopted safely.

Creating a human-like AI competitor is ‘stupid’

At an AI conference in Cannes, France in February, Russell was among the experts to initially dismiss OpenAI’s ChatGPT along with Turing award winner Yann LeCun. Russell also opined that artificial general intelligence (AGI) - when a system becomes so sophisticated it can do any intellectual task that humans can do - is far off.

Now Russell tells The Times that his AGI timeline has "shrunk a bit" to potentially arriving within 10 years, although he believes it is still "pretty unlikely."

At Cannes, Russell cautioned that “unless its only purpose is to be a benefit to humans, you are actually creating a competitor – and that would be obviously a stupid thing to do.”