AI needs a 'brake pedal', warns Anthropic co-founder

Artificial intelligence is rapidly advancing, prompting calls for mechanisms to ensure human control over its development. Jack Clark, a co-founder of AI company Anthropic, has voiced concerns that the technology is approaching a point where it could evolve without human oversight.

"You want the option to be able to take your foot off the gas and put your foot on the brake," Clark told BBC Newsnight. "Right now, it's like the AI industry has a gas pedal, but it doesn't have a brake pedal." He emphasized the necessity of government policy to maintain control over increasingly powerful AI systems that will have broad societal impacts.

“

You want the option to be able to take your foot off the gas and put your foot on the brake. Right now, it's like the AI industry has a gas pedal, but it doesn't have a brake pedal.

Clark suggested that society needs to develop new regulations to ensure confidence in AI systems, drawing a parallel with the establishment of regulatory frameworks for the oil industry at the turn of the last century. He believes a similar approach is needed to manage the benefits and risks associated with AI.

Despite these concerns, Anthropic recently welcomed a relatively hands-off executive order on AI from U.S. President Donald Trump. Major AI firms, including Anthropic, OpenAI, and Google, have not announced pauses in their research. Anthropic, valued at nearly $1 trillion by private investors, aims to be transparent about the capabilities and implications of AI technology, not just for commercial reasons, but to inform the public about the unusual nature of this developing field.

“

The world needs to do some thinking and we need to eventually develop some some new regulations that allow us to be confident in these systems.