Updated

Anthropic CEO Dario Amodei said China presents the 'toughest dilemma' for his proposed industry-wide AI slowdown, an initiative aimed at pausing or moderating the pace of frontier AI development amid safety concerns. OpenAI's Sam Altman, Google DeepMind chair Demis Hassabis, and other industry leaders have weighed in on the proposal as debate over AI safety governance intensifies. Amodei and other prominent voices have also begun warning about a related concept known as recursive self-improvement, in which AI models could begin building on themselves without human involvement, a prospect they say heightens the urgency of coordinated industry action. The debate has drawn a public rebuke from President Trump, who used social media and a phone call to Nvidia CEO Jensen Huang at the All-In Summit to dismiss opposition to AI data center buildouts as a 'hoax,' pushing back directly against Amodei's call for a slower pace of development. The clash between the administration's full-speed-ahead posture and safety-focused voices within the AI industry underscores the widening gap between commercial and political incentives on one side and precautionary risk concerns on the other.