US AI Policy Debate Escalates: Meta CEO Rejects Chinese Model Ban, Court Permits Book‑Destruction Training

Meta chief executive Mark Zuckerberg told the Financial Times that a U.S. prohibition on Chinese artificial‑intelligence models would not be an effective way to gain a competitive edge. He urged the United States to focus on removing bottlenecks and expanding open‑source AI development rather than banning models from firms such as DeepSeek, Qwen and Kimi, warning that excessive restrictions could concentrate power in a few large players.

In a separate development, a U.S. federal judge ruled that AI companies may legally purchase physical books, destroy the copies after digitising them, and use the resulting digital files to train language models. The decision, issued in the case Bartz v. Anthropic PBC, affirmed the practice as fair use provided the books are bought lawfully. The ruling follows a $1.5 billion settlement with Anthropic over alleged use of pirated works and has sparked a surge in the acquisition of rare and out‑of‑print books by technology firms, raising concerns in the publishing sector about cultural preservation.

Both events highlight a growing clash between U.S. regulatory approaches and the rapid expansion of AI capabilities, with industry leaders pushing for more open development while policymakers consider tighter controls.