Anthropic Denies It Could Sabotage AI Tools During War
WIRED • • nation
Key Points:
- Anthropic asserts it cannot manipulate or disable its AI model Claude once deployed by the US military, countering claims from the Trump administration that it might interfere with military operations.
- The Pentagon has designated Anthropic a supply-chain risk, barring the Department of Defense and contractors from using its software, prompting Anthropic to file lawsuits challenging the ban’s constitutionality.
- Anthropic executives emphasize they lack backdoor access or kill switches for Claude and have offered contractual guarantees to avoid controlling or vetoing military decisions, but negotiations with the Defense Department have stalled.
- The Department of Defense is working with third-party cloud providers to prevent unilateral changes by Anthropic to Claude systems and argues the ban is necessary to avoid risks to critical military operations.
- The dispute
