Anthropic CEO Warns AI Race Could Spin Out Of Control Within A Year, Calls For Slowdown

Anthropic's CEO Dario Amodei is pushing for a deliberate slowdown in AI development, a stance that carries weight given the company's impending IPO. His concerns are rooted in the rapid progress driven by recursive self-improvement and incidents like the OpenAI bot attack on Hugging Face, which suggest a growing risk of AI systems operating beyond human control. The proposed external review team at Anthropic, with its independent publishing rights, is a notable step towards accountability, though the lack of named reviewers or a clear timeline for their appointment means its effectiveness remains to be seen. For Asia, this discussion around AI safety and control is critical. As regional governments and companies invest heavily in AI, the potential for a "persistent botnet" causing billions in damages within a year, as Amodei warns, underscores the need for robust cybersecurity frameworks and international collaboration on AI governance. The call for an exemption from trade laws to facilitate safety talks, alongside continued limits on AI chip sales to China, reflects a strategic push for the US to maintain a lead in AI development, which could impact the competitive landscape for Asian AI firms and researchers.