Warren is pressing the Pentagon on the decision to grant AI access to classified networks

Sen. Elizabeth Warren (D-Mass.) sent A letter Defense Secretary Pete Hegseth wrote Monday expressing concern about the Pentagon’s decision to give Elon Musk’s company XAI access to classified networks.

“Grok, the controversial AI model developed by xAI, has provided disturbing results to users, including advising users on how to commit murders and terrorist attacks, generating anti-Semitic content, and creating child sexual abuse material,” the letter said.

Warren said Grok’s “apparent lack of adequate guardrails” could pose “serious risks to the safety of U.S. military personnel and the cybersecurity of classified systems.” Hegseth requested information about how the Department of Defense plans to “mitigate these potential national security risks.”

Warren wasn’t the first to express alarm that Grok, xAI’s controversial chatbot, gained access to secret systems. Last month, A The coalition of non-profit organizations urged The government immediately suspended Grok’s publication Federal agenciesincluding the Department of Defense, after X users repeatedly asked the chatbot to turn real photos of women, and in some cases of children, into real ones. Sexual images Without their consent. On the same day Warren sent her letter, a class-action lawsuit was filed against XAI alleging that Grok had created sexual content from real photos of the plaintiffs as minors.

The letter comes in the wake of the Pentagon’s decision to do so Anthropic label for supply chain risk After the AI ​​company refused to give the military unrestricted access to its AI systems. Anthropic was, until recently, the only AI company with systems in place for secret classification. In the midst of this conflict, the Ministry of Defense signed a memorandum Agreement with OpenAI In addition to XAI to use the two companies’ artificial intelligence systems in underground networks, according to Axios.

A senior Pentagon official confirmed that Grock was on board the plane for use in a secret location, but has not yet been used.

“It is unclear what assurances or documentation xAI provided to DoD regarding Grok’s security assurances, data handling practices, or safety controls, and whether DoD evaluated those assurances before allowing Grok access to the classified system,” Warren wrote.

Warren requested a copy of the deal reportedly reached between the Department of Defense and xAI regarding the use of Grok in classified systems and an explanation of how the department plans to ensure that Grok is not vulnerable to cyberattacks and “is not leaking sensitive or classified military information.”

(Last week, it was reported that a former employee at Musk’s government efficiency department to steal Americans’ personal data from the Social Security Administration and stored on a thumb drive – the latest accusation of Data leak related to DOGE.)

The Pentagon’s chief spokesman, Sean Parnell, said the department “looks forward to deploying Grok on its official artificial intelligence platform, GenAI.mil, in the very near future.”

GenAI.mil It is the Army’s secure enterprise platform for generative AI that gives DoD personnel access to large language models (LLMs) and other AI tools within government-approved cloud environments. It is designed to assist with primarily non-confidential tasks such as research, document drafting, and data analysis.