Anthropic’s risk-forward stance fuels debate over AI export bans

FT analysis shows Anthropic’s 2026 communications leaned heavily on risk, safety and regulation compared with OpenAI, a pattern some critics say helped spur Washington’s export ban on its Mythos and Fable models; while Anthropic positions itself as AI’s conscience, its tone has softened since 2023 as the technology gains prominence, and the export restrictions underscore the broader push to regulate frontier AI amid security concerns.