DOJ Won't Pledge No Further Penalties on Anthropic
SkimNews Take
The administration's refusal to foreclose further penalties means Anthropic faces rolling exposure: winning the lawsuit resolves today's pressure but leaves tomorrow's sanctions intact.
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- DOJ attorney James Harlow told US district judge Rita Lin he was "not prepared to offer any commitments" that the Trump administration would not levy additional penalties on Anthropic.
- President Trump is finalizing an executive order that would formally ban usage of Anthropic tools across the federal government, according to a person familiar with the matter (Axios first reported the plan).
- Anthropic filed two federal lawsuits Monday alleging the government unconstitutionally designated it a "supply-chain risk," with billions in revenue at stake as current and prospective customers drop out of deals.
- Judge Rita Lin moved the preliminary hearing date up to March 24 in San Francisco — later than Anthropic wanted but earlier than the April date the government proposed, citing the need for "a full record."
- The dispute began when Anthropic refused to let the military use its technology for "any lawful purpose," fearing broad surveillance of Americans and autonomous missile launches; the Pentagon says usage decisions are its prerogative.
- OpenAI and Google are moving forward with Pentagon deals to supplant Anthropic, despite employee pressure to push back on government demands over how their technology should be used.
- Legal scholars Harold Hongju Koh (Yale) and David Super (Georgetown) characterized the sanctions as part of a pattern of repurposing procurement law to punish perceived political enemies, with Super calling the "sabotage" label an "absurd stretch of the English language."
Why it matters: Anthropic has billions in revenue at risk as customers walk away, while OpenAI and Google absorb the Pentagon contracts the company refused. Legal experts warn the government's "multifront attack" — a pending executive order on top of the supply-chain-risk label — could deter other AI contractors from setting ethical boundaries on military use, even if Anthropic ultimately wins in court.
