Google expands Pentagon AI access after Anthropic declines
▼ Summary
– Google has granted the U.S. Department of Defense access to its AI for classified networks, allowing all lawful uses.
– The deal follows Anthropic’s refusal to grant the DoD unrestricted AI access due to concerns about domestic mass surveillance and autonomous weapons.
– In response, the DoD labeled Anthropic a “supply-chain risk,” and the two are now in a lawsuit, with a judge granting Anthropic an injunction.
– Google, OpenAI, and xAI have all signed deals with the DoD, with Google’s contract including non-binding language against mass surveillance and autonomous weapons.
– Google entered the deal despite 950 employees signing an open letter urging it to follow Anthropic’s lead and require similar guardrails.
Google has expanded access to its artificial intelligence systems for the U.S. Department of Defense, granting the Pentagon permission to use the technology on classified networks for all lawful purposes, according to multiple news outlets.
The move follows a very public clash between Anthropic and the Trump administration. The AI safety company refused to offer the Pentagon the same unrestricted terms, insisting on guardrails to prevent its models from being deployed in domestic mass surveillance or autonomous weapons. In response, the DoD labeled Anthropic a “supply-chain risk,” a designation typically reserved for foreign adversaries. The two sides are now locked in a legal battle, with a judge recently granting Anthropic an injunction against that label while the case proceeds.
Google is the third major AI firm to capitalize on Anthropic’s fallout. OpenAI and xAI quickly signed their own agreements with the DoD after Anthropic’s refusal. According to The Wall Street Journal, Google’s contract includes language stating the company does not intend for its AI to be used in domestic mass surveillance or autonomous weapons, a clause similar to what OpenAI secured. However, the WSJ notes it remains unclear whether those provisions are legally binding or enforceable.
This deal comes despite internal dissent at Google. More than 950 employees have signed an open letter urging the company to follow Anthropic’s lead and refuse to sell AI to the Defense Department without enforceable safeguards. Google did not respond to a request for comment on the matter.
(Source: TechCrunch)
