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

Anthropic co-founder Jack Clark called for urgent regulatory measures to control AI advancement, warning that the technology risks developing beyond human oversight. In an interview with BBC Newsnight, Clark compared the situation to the oil boom, where society later imposed policies to manage risks and ensure public trust. He noted that AI systems like Anthropic’s Claude—currently 80% self-generated code—could achieve full autonomy in as little as two years, raising concerns about economic disruption and job losses from AI-driven automation. Clark argued that governments must establish frameworks to regulate AI, preventing unchecked progress that could harm society. He dismissed recent executive orders, such as the one issued by President Donald Trump, as insufficient, as they lack mandatory safety testing requirements. Major AI firms, including Anthropic, OpenAI, and Google, have not committed to pausing research despite growing public fears about AI’s capabilities. Anthropic, founded five years ago by CEO Dario Amodei, is preparing for a high-profile public stock listing, with a valuation nearing $1 trillion. Clark insisted the company’s warnings about AI risks are not for public relations but reflect genuine concerns about the technology’s trajectory. Anthropic has previously clashed with the U.S. Department of Defense over AI’s potential misuse in surveillance and autonomous warfare. Clark expressed alarm for future generations, stating that unchecked AI could disrupt economies by replacing jobs, though creative roles may remain resilient. He highlighted recent tech layoffs, often attributed to AI’s ability to perform tasks previously handled by engineers. Despite warnings, no major AI company has voluntarily halted research, leaving regulatory gaps that Clark says must be addressed before irreversible consequences arise.