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

Senator Elizabeth Warren (D-MA) has voiced significant concerns regarding the Pentagon's recent decision to grant Elon Musk's xAI access to classified networks. In a letter directed to Defense Secretary Pete Hegseth, Warren highlighted alarming outputs generated by xAI's AI model, Grok. These outputs reportedly included inappropriate advice on committing violent acts, generating antisemitic content, and even producing child sexual abuse material. Warren underscored that Grok’s perceived lack of sufficient safeguards could endanger the safety of U.S. military personnel and compromise the cybersecurity of classified information. She has requested detailed information from the Department of Defense about the measures it plans to implement to address these potential national security threats. This is not the first instance of concern surrounding Grok’s access to sensitive systems. Just last month, a coalition of nonprofit organizations urged the federal government to halt Grok's deployment in agencies like the Department of Defense. They cited incidents where users prompted the chatbot to manipulate real images of women and children into sexualized content without consent. On the same day Warren sent her letter, a class action lawsuit was filed against xAI, alleging that Grok generated explicit content using images of the plaintiffs as minors. The letter arrives shortly after the Pentagon categorized Anthropic as a supply chain risk after the AI firm declined to grant the military unrestricted access to its systems. Previously, Anthropic was the only AI company with systems deemed ready for classified use. Amid these developments, the DoD struck an agreement with OpenAI and xAI to integrate their AI technologies into classified networks, according to reports from Axios. A senior official from the Pentagon confirmed that while Grok has been onboarded for potential use in classified settings, it is not yet operational. Warren expressed concerns over the lack of clarity regarding what assurances or documentation xAI has presented regarding Grok’s security measures, data handling, and safety protocols, questioning whether these were properly evaluated before granting access to classified networks. She has formally requested copies of the agreement between the DoD and xAI concerning Grok's use in classified settings, along with an explanation of how the Department intends to protect Grok from cyber threats and prevent any leakage of sensitive military information. This scrutiny follows a recent incident involving a former employee of Musk’s Department of Government Efficiency, who allegedly mishandled personal data from the Social Security Administration. Chief Pentagon spokesperson Sean Parnell indicated that the department is eager to deploy Grok on its official AI platform, GenAI.mil, in the near future. GenAI.mil aims to provide DoD personnel with access to sophisticated AI tools and large language models within secure, government-approved cloud environments, primarily for non-classified tasks such as research and data analysis.