Warren Demands Answers After Pentagon Grants xAI Access to Classified Networks Amid Safety Concerns

Warren presses Pentagon over decision to grant xAI access to classified networks | TechCrunch

Sen. Elizabeth Warren (D-MA) sent a letter to Defense Secretary Pete Hegseth on Monday expressing concern over the Pentagon's reported decision to give Elon Musk's company xAI access to classified networks. "Grok, the controversial AI model developed by xAI, has provided disturbing outputs for users, including giving users 'advice on how to commit murders and terrorist attacks,' generating antisemitic content, and creating child sexual abuse material," the letter reads. Warren said Grok's "apparent lack of adequate guardrails" could pose "serious risks to the safety of U.S. military personnel and to the cybersecurity of classified systems." She demanded Hegseth provide information on how the Department of Defense plans to "mitigate these potential national security risks." Warren isn't the first to express alarm at Grok, xAI's controversial chatbot, gaining access to classified systems. Last month, a coalition of nonprofits urged the government to immediately suspend the deployment of Grok in federal agencies, including the DoD, after X users repeatedly prompted the chatbot to turn real photos of women, and in some cases children, into sexualized images without their consent. The same day Warren sent her letter, a class action lawsuit was filed against xAI alleging Grok had generated sexual content from real images of the plaintiffs as minors. The letter comes in the aftermath of the Pentagon's decision to label Anthropic a supply chain risk after the AI firm refused to give the military unrestricted access to its AI systems. Anthropic had been, until recently, the only AI company with classified-ready systems. In the midst of that conflict, the DoD signed an agreement with OpenAI as well as xAI to use the two companies' AI systems in classified networks, according to Axios. A senior Pentagon official confirmed that Grok was onboarded to be used in a classified setting, but is not yet being used. "It is unclear what assurances or documentation xAI has provided to the Department of Defense about Grok's security safeguards, data-handling practices, or safety controls, and whether DoD has evaluated those assurances before reportedly allowing Grok access to classified system," Warren writes. Warren requested a copy of the deal reportedly reached between the DoD and xAI on the use of Grok in classified systems and an explanation of how the department plans to ensure Grok is not exposed to cyberattacks and will "not leak sensitive or classified military information." (Last week, a former employee of Musk's Department of Government Efficiency reportedly stole Americans' personal data from the Social Security Administration and stored it on a thumbdrive -- the latest accusation of DOGE-related data leakage.) Chief Pentagon spokesperson Sean Parnell said the department "looks forward to deploying Grok to its official AI platform GenAI.mil in the very near future." GenAI.mil is the military's secure enterprise platform for generative AI that gives DoD workers access to large language models (LLMs) and other AI tools within government-approved cloud environments. It is designed to help with primarily non-classified tasks like research, document drafting, and data analysis.

Warren demands Hegseth share information about xAI's access to classified networks

Sen. Elizabeth Warren, D-Mass., sent a letter to Defense Secretary Pete Hegseth on Monday demanding information about the Pentagon's reported decision to give Elon Musk's xAI access to classified networks. In the letter, Warren raised concerns that xAI and its AI-powered chatbot, Grok, could present critical safety and reliability risks to America's military. Warren cited the possibility that Grok could leak classified information to adversaries, be manipulated based on biased or inaccurate data, or lack critical safety controls and put service members in danger. "I am concerned that Grok's apparent lack of adequate guardrails could pose serious risks to the safety of U.S. military personnel and to the cybersecurity of classified systems," Warren wrote in the letter, seen by NBC News, "especially if Grok is given sensitive military information and access to operational systems." "It is unclear what assurances or documentation xAI has provided to the Department of Defense about Grok's security safeguards, data-handling practices, or safety controls, and whether DoD has evaluated those assurances before reportedly allowing Grok access to classified systems," Warren wrote. In late February, the Pentagon and xAI reached a deal that would pave the way for xAI's systems to be used on classified networks, according to Axios, citing an anonymous Defense Department official. The announcement came in the middle of the Pentagon's rupture with rival AI company Anthropic, which had insisted on stronger guarantees that the Pentagon would not use its AI systems for domestic surveillance or direct use in deadly weapons. It is unclear whether xAI's reported deal with the Pentagon allows for a wider range of uses than Anthropic's proposed contract. In her letter, Warren requested a full copy of the agreement reportedly reached between the Defense Department and xAI. In July, xAI received a contract worth up to $200 million from the Pentagon to develop new AI applications for the Defense Department. In a statement, the military's Chief Digital and Artificial Intelligence Office said the deal would "broaden DoD use of and experience in frontier AI capabilities and increase the ability of these companies to understand and address critical national security needs." In recent months, xAI has come under intense scrutiny at several government agencies, according to The Wall Street Journal, due to concerns about the safety and reliability of its Grok services. The Journal reported that the Pentagon had previously raised concerns about Grok and "questioned whether it was aligned with government ethics and standards." In late December and early January, xAI received widespread attention for Grok's ability to manipulate images of people, primarily women and children, to remove subjects' clothing. The phenomenon prompted California's Attorney General Rob Bonta to launch an investigation into the company, while countries including Indonesia and Malaysia banned Grok over the sexualized photos. Grok has been framed by Musk as being more permissive and less "woke" than AI models from competitors like Anthropic, OpenAI and Google. These lower guardrails have led the chatbot to generate controversial or incendiary outputs at times, for example churning out antisemitic posts after a technical update in July. Similar comments from Grok in November prompted the French government to open an inquiry into xAI. In February, the European Union's data privacy office launched an investigation into X, the social media platform closely associated with Grok and xAI, over concerns that X was spreading sexualized AI-generated images of women and children. Musk called similar moves in February from the Paris prosecutor's office "a political attack." Warren's letter asked Hegseth to share how the Defense Department plans to mitigate the risks of inaccurate responses and other safety shortcomings, along with all communications that led to the reported agreement.

Warren demands information on xAI access to classified Pentagon networks

Sen. Elizabeth Warren (D-Mass.) is demanding more information from Defense Secretary Pete Hegseth on xAI's access to Pentagon classified networks. She also asked him to address worries that the company's chatbot, Grok, could pose reliability concerns for the U.S. military. Warren, who sits on the Senate Armed Services Committee, said in a Sunday letter that she was concerned that Grok could present safety risks to armed forces, 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 the four-page letter. "It is unclear what assurances or documentation xAI has provided to the Department of Defense about Grok's security safeguards, data-handling practices, or safety controls, and whether DoD has evaluated those assurances before reportedly allowing Grok access to classified systems." Chief Pentagon spokesperson Sean Parnell said the Department of Defense (DOD) said no one single prompt is used "to define the operational capability and wanting to work with the government." "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," he said. DOD and xAI struck a deal last month to bring its AI models to the military's classified systems, an agreement that came as the Pentagon was publicly feuding with Anthropic, which demanded that Claude, its AI model, not be used in firing autonomous weapons and domestic surveillance. Warren, in her letter, asked for a full copy of the agreement between xAI and the Pentagon. Last month, The Wall Street Journal reported that officials in multiple government agencies, including the National Security Agency and General Services Administration, raised concerns over Grok's safety and reliability. Musk's xAI got a nearly $200 million Pentagon contract last July to develop an AI application for the DOD. "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," Warren asked in the letter.