Zuckerberg expanded his critique of AI regulation, arguing open development better supports U.S. innovation than restrictions on Chinese models or tighter control by a handful of frontier labs.
Meta CEO Mark Zuckerberg said the U.S. government should not block Chinese AI models as it competes with China in artificial intelligence, arguing that bans are not "an effective solution" and that Washington should instead address domestic bottlenecks. In rare interviews with the Financial Times and the New York Times, Zuckerberg also criticized efforts by leading AI companies to shape regulation in ways that could concentrate power and curb innovation, while reiterating his support for more open AI development. His comments came as Chinese startup Moonshot AI's Kimi K3 drew scrutiny from U.S. officials over allegations it was trained on American AI data using distillation techniques and as the broader industry remains split between advocates of open model development and backers of tighter controls on frontier systems. Zuckerberg challenged the "alignment risk" argument advanced by some rivals, saying a single universally aligned superintelligence is unrealistic and arguing instead for more personalized AI tools. The debate has intensified as more than 1,100 employees from companies including OpenAI, Anthropic, Google and Meta backed a petition for an international mechanism that could slow frontier AI development when necessary.