Ex‐OpenAI researcher warns AI could kill humans, Anthropic safety executive backs grim forecast

The departure of Jacob Coxon from the AI industry and the stark warnings from Anthropic's Evan Hubinger about AI's existential risk, estimated at over 10 percent in the next decade, underscore a growing internal conflict within leading US AI firms. This is not merely about individual researchers' fears; it reflects a systemic tension between rapid development and safety, even as companies like Anthropic acknowledge the risks but feel compelled to race ahead due to competitive pressures. For Asia, this internal debate in the US points to the critical need for regional governments and AI developers to proactively establish clear regulatory frameworks. Without federal AI regulation in the United States, the competitive push for recursive self-improvement could create a fragmented global safety landscape. Asian nations, particularly those investing heavily in AI research and deployment, must consider how to balance innovation with robust safety protocols to avoid inheriting unmitigated risks from an unregulated global race. The calls for international coordination on AI development, as voiced by OpenAI's chief scientist Jakub Pachocki, resonate strongly for Asian policymakers aiming to shape a secure AI future.