Amazon's Jassy Triggered Anthropic Model Export Ban
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Andy Jassy told Treasury Secretary Scott Bessent and other officials that Amazon researchers used Anthropic's Claude Fable 5 to extract information usable in cyberattacks, according to the Wall Street Journal.
- Anthropic cut off worldwide access to the Fable 5 and Mythos 5 models on Friday after the government imposed an export control ban.
- Amazon acknowledged governments routinely seek its counsel on security risks but declined to detail the discussions; AWS was also affected by the model cutoff.
- David Sacks, Trump's former AI czar, said a "highly credible trusted partner" of both Anthropic and the USG reported a jailbreak, and claimed the administration asked CEO Dario Amodei to fix it or de-deploy — which he refused.
- Anthropic countered in a blog post that the capabilities triggering government concern are already available in other publicly accessible models.
- The Information and Reuters corroborated that Amazon — a major Anthropic investor — communicated security concerns to the government.
Why it matters: Amazon is a major Anthropic investor, yet its CEO's reported tip-off triggered export controls on a company in which Amazon holds a stake — an extraordinary investor-versus-porticiary conflict. Anthropic's pushback that the same capabilities exist in competing models underscores that the restrictions may limit access to these specific models without actually constraining the underlying capability.
