Elizabeth Warren demands Pentagon explain xAI's access to classified networks amid security concerns

Elizabeth Warren Challenges Pentagon on Grok AI Model Security

Senator Elizabeth Warren has sent a pointed letter to Defense Secretary Pete Hegseth demanding comprehensive information about the Pentagon's decision to grant Elon Musk's xAI access to classified networks 1

1

4

4

. The Massachusetts Democrat, who sits on the Senate Armed Services Committee, raised alarm about potential security risks posed by xAI's Grok AI model, citing its controversial history of generating disturbing outputs including instructions on committing murders and terrorist attacks, antisemitic content, and child sexual abuse material 2

2

.

"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 her four-page letter 4

4

. The senator specifically questioned whether the Defense Department had properly evaluated xAI's security safeguards, data-handling practices, and safety controls before granting access to sensitive military information.

National Security Agencies Raised Red Flags

Warren's letter reveals that multiple federal agencies, including the National Security Agency and the General Services Administration, had raised concerns about Grok's reliability before the Pentagon moved forward with the agreement 3

3

. The NSA conducted a classified review and "determined Grok had particular security concerns that other models didn't," according to the senator's correspondence 3

3

. Government testing showed Grok is more susceptible than competing models to data poisoning attacks, where manipulated data corrupts the system's outputs—a serious vulnerability for a tool being considered for weapons development and battlefield intelligence.

"Were Grok to leak government information, this could reveal sensitive military plans, U.S. intelligence efforts, and potentially put service members in danger," Warren warned 5

5

. The timing of her inquiry carries additional weight, as it arrived the same day three Tennessee minors filed a federal class action lawsuit against xAI, alleging Grok generated child sexual abuse material based on their real photographs 3

3

.

Pentagon Deal Emerges Amid Anthropic Controversy

The Pentagon and xAI reached a deal in late February that would enable xAI's systems to be used on classified networks, according to reports 4

4

. This agreement surfaced during the Pentagon's public rupture with Anthropic, which had insisted on stronger guarantees that the Defense Department would not use its AI systems for domestic surveillance or direct deployment in autonomous weapons 4

4

. The Defense Department subsequently designated Anthropic as a supply-chain risk after the AI lab refused to make concessions on how its technology could be used by the military 1

1

.

Warren characterized the Pentagon's treatment of Anthropic as "retaliation," arguing in a separate letter that "the DoD is trying to strong-arm American companies into providing the Department with the tools to spy on American citizens and deploy fully autonomous weapons without adequate safeguards" 1

1

. Anthropic had been the only AI company with classified-ready systems before the controversy, with its Claude model deployed in real military operations. Following the designation, both OpenAI and xAI were announced as replacements 3

3

.

Questions About Ethical Considerations and Safeguards

xAI received a contract worth up to $200 million from the Pentagon in July to develop new AI applications for the Defense Department 4

4

. However, Warren's letter highlights the absence of public information about what assurances xAI provided regarding Grok's capabilities and limitations. "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," the senator wrote 2

2

.

There are no records of xAI questioning the reach of the Pentagon's "all lawful purposes" standard, which Anthropic had contested 3

3

. Warren requested a full copy of the agreement between xAI and the Pentagon, along with all internal communications about the deal, asking whether any testing or evaluation took place before access was granted 3

3

. One of her ten questions asks directly whether safeguards exist to ensure Grok does not cause "erroneous targeting decisions" if deployed in critical operational systems.

Pentagon Defends AI Deployment Plans

Chief Pentagon spokesperson Sean Parnell defended the decision, stating 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 Defense Department workers access to large language models and other AI tools within government-approved cloud environments, designed primarily for non-classified tasks like research, document drafting, and data analysis . A senior Pentagon official confirmed that Grok was onboarded to be used in a classified setting but is not yet being deployed 2

2

.

Warren has requested responses by March 30, as concerns mount about the leakage of classified information and the broader implications of AI deployment in military contexts without transparent oversight. The controversy highlights growing tensions between rapid AI adoption for national security purposes and the need for robust ethical considerations around surveillance, autonomous weapons, and military personnel safety.