Pentagon faces scrutiny over xAI’s Grok access

On Monday, Senator Elizabeth Warren sent a letter to Defense Secretary Pete Hegseth expressing concern over the decision by the Pentagon (the U.S. Department of Defense) to give xAI (Elon Musk’s AI company) access to classified networks. In her letter, Warren raised concerns over the safety record of Grok, the AI model developed by xAI. Senator Elizabeth Warren wrote, “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,”. Warren argued that the model’s apparent lack of adequate guardrails could pose risks to the safety of U.S. military personnel and cybersecurity. She also questioned what assurances or documentation xAI has provided to the Department of Defense regarding Grok’s security safeguards, data-handling practices, or safety controls, and whether the department evaluated those assurances before reportedly allowing the model access to classified systems.

This controversy arrives as the Department of Defense navigates a complex procurement landscape. The Pentagon recently labeled Anthropic a supply chain risk after the firm refused to give the military unrestricted access to its AI systems. Following that dispute, the Department of Defense signed an agreement with OpenAI and xAI to use their 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. Despite this, chief Pentagon spokesperson Sean Parnell stated that the department looks forward to deploying Grok to its official AI platform GenAI.mil (the military’s secure enterprise platform for generative AI) in the very near future. However, this stance faces external pressure. Last month, a coalition of nonprofits urged the government to suspend the deployment of Grok in federal agencies. Additionally, a class action lawsuit was filed against xAI alleging that Grok generated sexual content from real images of the plaintiffs as minors.

The debate over xAI’s access also coincides with concerns regarding data security. Last week, a former employee of the Department of Government Efficiency (DOGE) reportedly stole Americans’ personal data from the Social Security Administration.

Why it matters

The Pentagon’s integration of private AI models into classified systems is creating friction, as lawmakers and safety advocates question whether these firms can meet military-grade security standards.