U.S. Weighs Limiting AI Exports to China, Risking Conflict and Slowing Innovation

The U.S. is considering limiting exports of AI systems to China over concerns they could develop dangerous capabilities, risking greater conflict.

Tech companies like Microsoft and Google advocated restricting "frontier models" in an influential white paper. This could help their business models.

Export controls could reduce collaboration between AI researchers in the U.S. and China, slowing innovation.

Technical challenges make restricting frontier models difficult, as models can easily be recreated. Companies are already finding workarounds to chip export limits.

Some experts argue frontier model concerns distract from regulating present harms of AI like privacy, jobs, and the environment.