AI needs a 'brake pedal', warns Anthropic co-founder
Anthropic co-founder Jack Clark calls for the ability to slow AI progression, warning it may soon develop without human input.

As artificial intelligence continues to advance at a rapid pace, Anthropic co-founder Jack Clark is sounding the alarm, calling for the development of a 'brake pedal' to slow down the technology's progression. In a recent interview with BBC Newsnight, Clark warned that the world is nearing a point where AI could develop without human input, highlighting the need for a mechanism to pause or slow down AI research and development. Clark drew a parallel between the AI industry and the oil boom of the turn of the last century, suggesting that society needs to develop a regulatory framework that allows people to be confident in AI systems.
'You want the option to be able to take your foot off the gas and put your foot on the brake,' Clark said. 'Right now, it's like the AI industry has a gas pedal, but it doesn't have a brake pedal.' The push for a 'brake pedal' comes as Anthropic's popular chatbot, Claude, operates on code that is 80% self-written, with the potential to reach 100% within two years. Clark stressed that this would have huge implications, emphasizing the need for people to keep control of AI systems through government policy.
Anthropic has welcomed a recent executive order on AI from US President Donald Trump, which takes a relatively hands-off approach to regulating the industry. However, Clark did not outline how a 'brake pedal' for AI research and development could be created, instead emphasizing the need for a sensible policy and regulatory framework. The conversation around AI regulation comes as Anthropic prepares to debut on the public stock market, with a valuation estimated to be nearly $1tn (£745bn).
Clark emphasized that the company's motivation for publicly discussing AI's growing capabilities is not to burnish its reputation, but to 'tell the world what we're seeing inside these companies with this unusual technology.' As AI continues to advance, Clark expressed concerns about the potential risks, including disruption to the economy and the potential for AI to take over certain jobs. However, he also suggested that individuals with creative skills and a liberal arts education may have an advantage in an AI-driven economy, encouraging young people to 'develop a hobby' and pursue their interests.
Source: BBC Technology