Anthropic controversy raises the question of who can block dangerous AI

The Verge’s latest Decoder discussion centers on a hard governance question: who should decide when an AI system is too dangerous to release or keep available?

The episode connects Anthropic, the Trump administration, and the Fable 5 controversy, illustrating how quickly model decisions can become political and regulatory flashpoints.

For AI companies, the issue is not only technical risk assessment. It is also legitimacy: who has authority, what evidence is public, and how much government pressure should shape access to powerful models.