The Pentagon faced a competitive contractor, Anthropic, on ethical restrictions on its technology
U.S. Secretary of War Pete Hegseth speaks during a visit to Sierra Space in Louisville, Colorado, February 23, 2026. © Getty Images / AAron Ontiveroz / The Denver Post; Samuel Boivin / NurPhoto
U.S. War Department reportedly signed an agreement with xAI Elona Muska to integrate the chatbot A grout with secret military systems, expanding force on a competitive supplier, Anthropic, who refuses to lift Claude's safety features.
The agreement, first notified by the fresh York Times and confirmed by Axios on Monday, would make Grok the second artificial intelligence strategy approved for usage in the most delicate military networks, where intelligence analyses, weapons improvement and combat operations are conducted.
So far Claude of Anthropic was the only model available on secret platforms, thanks to a partnership with Palantir Technologies.
The agreement was concluded at a time erstwhile War Secretary Pete Hegseth called on the president of Anthropic, Dario Amodeia, to a gathering that, as sources expect, would be strained on Tuesday at the Pentagon.
According to Axios, Hegseth is to present an ultimatum: agree to make Claude available ‘for all legitimate purposes’ without additional safeguards, otherwise they endanger to have consequences, including possible designation of "Provision chain threat" – a word usually reserved for entities associated with abroad opponents.
Anthropic opposes the Pentagon's demands to abolish restrictions preventing its technology from being utilized for mass surveillance of Americans or full autonomous weapons systems without human involvement.
XAI reportedly agreed to demands, but so far has not commented on these reports.
According to people informed about the calls, Google is besides ‘close’ conclude a contract allowing the secret usage of the Gemini model while OpenAI continues ‘not close’, due to the fact that it continues to work on safety technology.
Pentagon officials admit that replacing Anthropic in secret systems can origin short-term disruption.
This model was utilized during the Venezuelan president Nicolas Maduro's capture operation last period – it was the first known case of direct function of artificial intelligence in an active military raid.
Anthropic positions itself as a safety conscious alternate in the artificial intelligence industry.
President Amodei has repeatedly warned against existential threats posed by unlimited artificial intelligence, including ‘risk of autonomy’.
The head of the company's safety investigation team, Mrinank Sharma, abruptly resigned last week, announcing a mysterious informing that "The planet is in danger."
Translated by Google Translator
source:https://www.rt.com/news/632949-pentagon-grok-ai-contract/
