Michaela
U.S. Senator Elizabеth Warren has raised concerns with Defense Secretary Pete Hegseth over reports that the Department of Defense (DoD) may grant Elon Musk’s AI firm xAI access to classified systems, despite warnings from multiple federal agencies.
Key Points
- Senator Elizabeth Warren formally questioned the DoD over allowing xAI’s Grok access to classified systems.
- Grok has shown unsafe behavior, including generating violent and inappropriate content, and prior chat leaks raised privacy concerns.
- Warren highlighted cybersecurity vulnerabilities like susceptibility to data poisoning and possible exposure to adversaries.
In a March 15 letter, Warren said that both the National Security Agency (NSA) and the General Services Administration (GSA) had raised concerns about the reported move. She also pointed to troubling behavior from xAI’s chatbot, Grok, citing instances where it allegedly produced harmful outputs, including guidance related to violent acts, antisemitic material, and inappropriate content involving minors.
“According to recent reports, the National Security Agency “conducted a classified review . . . [and] determined Grok had particular security concerns that other models . . . didn’t,” Warren wrote. “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, especially if Grok is given sensitive military information and access to operational systems,” Warren continued.
Senator Warren called on Hegseth to provide immediate details on how the DoD intends to address potential national security risks tied to the use of xAI’s Grok system. She also raised concerns that the company may not have implemented sufficient safeguards, pointing to reports that large volumes of supposedly private Grok conversations were previously accessible through search engines.
Related: ChatGPT’s ‘Adult Mode’ Experiences Further Delay at OpenAI
Warren further cited findings reported by The Wall Street Journal, which indicated that individuals familiar with government testing observed the chatbot to be more vulnerable than comparable models to “data poisoning,” a process in which manipulated or misleading inputs compromise underlying datasets. The senator argued that his susceptibility could increase the risk of exploitation by hostile actors targeting U.S. systems.
“As Secretary of Defense, you are responsible for protecting highly sensitive and classified information and procuring the best tools through a competitive aсquisition process. But under your leadership, the Department awarded xAI a contract worth up to $200 million under circumstances that have raised questions about the process for including xAI, because “xAI didn’t have the kind of reputation or track record that typically leads to lucrative government contracts,” Warren wrote.
Related: OpenAI Debated Police Call Before Canada Mass Shooting Suspect Chats Online
Additionally, Warren warned that any leak of government data by xAI’s Grok could expose sensitive military operations, intelligence activities, and potentially endanger service members. She emphasized that it remains unclear what security protocols, dаta-handling measures, or safety assurances xAI has provided to the DoD, and whether these safeguards were reviewed before granting Grok access to classified networks.
Frequently Asked Questions
MICHAELA
Michaela is a news writer focused on cryptocurrency and blockchain topics. She prioritizes rigorous research and accuracy to uncover interesting angles and ensure engaging reporting. A lifelong book lover, she applies her passion for reading to deeply explore the constantly evolving crypto world.
