Anthropic Urges AI Slowdown Amid Fears of Unchecked Progress

Anthropic, the AI company behind the assistant Claude, has called for a slowdown in the race to develop advanced AI systems. The proposal comes as firms pour hundreds of billions into AI infrastructure, raising concerns about unchecked progress. According to the company, rapid advancements—especially in self-improving models—demand a more cautious approach. The idea stems from growing worries over AI’s ability to evolve without full human oversight. Anthropic’s recent Claude Mythos Preview software even uncovered long-hidden security flaws in existing programs, highlighting potential risks. Yet the company admits that reaching an industry-wide agreement on a pause would be difficult.

Anthropic opposes unilateral slowdowns, warning they could simply hand dominance to other players, particularly China. CEO Dario Amodei stressed the challenge of coordination in an AI arms race where nations and corporations compete fiercely. Instead, the firm wants to work with others to create a verification system, ensuring any moratorium is followed.

During this proposed slowdown, Anthropic plans to focus on adapting societal structures and advancing research. The goal is to ensure future AI systems align with humanity’s best interests. Their stance reflects a long-standing push for stricter industry regulation. The call for a pause arrives as AI development accelerates, with massive investments fuelling the race. Anthropic’s proposal includes monitoring mechanisms and a focus on safety research. Whether major developers will agree remains uncertain, but the company insists collaboration is essential to manage AI’s rapid growth.