AI needs a 'brake pedal' says Anthropic co-founder

Jack Clark, one of the founders of artificial intelligence company Anthropic, has said that AI systems are nearing the point where it no longer needs human input to develop and that we need the ability to slow down its development.

In an interview with BBC’s Newsnight, Clark, who serves as Head of Policy at Anthropic, said that the company’s chatbot Claude already runs on code 80% of which the system wrote itself. He estimated that the system would run entirely on self-generated code within two years from now.

This presents “huge implications” according to Clark, and he warned that humanity needs a way to slow down the progress of AI if needed. “You want the option to be able to take your foot off the gas and put your foot on the brake”, he told the BBC. “Right now, it's like the AI industry has a gas pedal, but it doesn't have a brake pedal.”

Clark did not give details about how this ‘brake pedal’ would work, or how it might be implemented, but he compared the situation to the oil boom at the start of the twentieth century. “Society's response was to come up with a sensible policy and regulatory framework that gave people confidence in oil and the benefits that oil could provide to the world, and meant that you didn't have to worry about the personalities of the people leading the companies”, he said. “That's clearly where we end up here.”

Anthropic is preparing to go public on the stock market, one of the first of the new generation of AI companies to do so. The growth of artificial intelligence since Anthropic’s founding five years ago is such that private investors value the company at almost $1 trillion.

Clark also spoke about the way that artificial intelligence could disrupt jobs and the future for young people. “I am worried for my kids if we as a society don't have a serious conversation about what the implications of AI's continued advances mean,” he said. “There are potentially great benefits. There are also risks.” He added that young people who are afraid of the advance of AI should “develop a hobby” and pursue education in the liberal arts.

Clark’s warnings come at the same time as a team at the University of Cambridge are using AI to create a “fundamentally new” type of vaccine, which could prevent against future outbreaks or pandemics. The new approach has been heralded as a “game changer” for vaccine development, with progress being made for vaccines for seasonal flu, coronaviruses, and Ebola. Some vaccines are now in human trials and Science Minister, Lord Vallance, said “this work could help speed up the roll out vaccines to benefit people all over the world for the long-term.”