Elizabeth Warren rips Pentagon for allowing xAI into top secret systems, revealing a nightmare lurking in the shadows

Senator Elizabeth Warren is demanding answers from Defense Secretary Pete Hegseth regarding xAI’s access to the Pentagon’s classified networks, raising some serious concerns about the company’s chatbot, Grok, as reported by The Hill. She also wants to know how the Department of Defense (DOD) plans to address potential reliability issues that Grok could pose for the U.S. military.

Warren, who sits on the Senate Armed Services Committee, laid out her worries in a letter. She’s particularly concerned that Grok could present significant safety risks to our armed forces, specifically citing the possibility of the chatbot leaking classified information.

“Were Grok to leak government information, this could reveal sensitive military plans, U.S. intelligence efforts, and potentially put service members in danger,” she wrote in the four-page letter. It’s still unclear what kind of assurances or documentation xAI has actually provided to the DOD about Grok’s security safeguards, its data-handling practices, or its safety controls. Warren wants to know if the DOD even evaluated those assurances before reportedly giving Grok access to classified systems.

It sounds like the Pentagon is all-in on this collaboration with xAI

Chief Pentagon spokesperson Sean Parnell, however, said that no single prompt is used “to define the operational capability and wanting to work with the government.” Parnell was pretty enthusiastic about the partnership, stating, “xAI’s Grok remains a competitive frontier model — no media spin can change that. The Department is excited to have xAI, one of America’s national champion frontier AI companies onboard and looks forward to deploying Grok to its official AI platform GenAI.mil in the very near future.”

This whole situation comes after the DOD and xAI struck a deal last month to bring its AI models into the military’s classified systems. This agreement actually followed some public friction between the Pentagon and Anthropic, another AI company. Anthropic had demanded that its AI model, Claude, not be used in firing autonomous weapons or for domestic surveillance. Warren, looking for full transparency, has specifically asked for a complete copy of the agreement between xAI and the Pentagon in her letter.

It’s not just Senator Warren who has concerns, either. Last month, a report indicated that officials in multiple government agencies, including the National Security Agency and the General Services Administration, also raised questions about Grok’s overall safety and reliability. This isn’t just a hypothetical issue; xAI actually secured a substantial contract last July, nearly $200 million, from the Pentagon to develop an AI application specifically for the DOD.

Warren’s letter also pressed for details on the protective measures in place. She asked, “What, if any, safeguards are in place, both in the agreement and within the Department writ large, to ensure that Grok is not exposed to cyberattacks, including data poisoning attacks, that could compromise its outputs.”