Nonprofits urge OMB to pause Grok in federal agencies over safety and deepfake risks

A coalition of consumer and digital-rights nonprofits is pushing the US government to halt deployments of Grok, arguing the xAI chatbot’s safety failures are incompatible with federal risk standards—especially as agencies consider using it for sensitive work.

Federal Grok deployment faces new compliance pressure

The groups’ open letter calls on the Office of Management and Budget (OMB) to immediately pause Grok’s rollout, formally investigate whether proper oversight occurred, and clarify whether Grok meets the administration’s “truth-seeking and neutral” requirements for AI systems. The letter cites recent incidents where users prompted Grok to generate sexualized images of real people, including minors, without consent. Bloomberg reported the system generated thousands of “undressed” images per hour that were then spread on X at scale, intensifying concerns about abuse pathways and content controls in image generation workflows (Bloomberg).

The letter also points to the newly enacted Take It Down Act (a federal law targeting revenge porn and explicit deepfakes) as raising the stakes for government vendors handling content generation and moderation (White House).

National security and procurement risks for LLM vendors

xAI previously secured a path into federal procurement: an agreement with the General Services Administration and a separate Department of Defense contract that can reach 200 million dollars, according to Axios (Axios). In mid-January, Defense Secretary Pete Hegseth said Grok would operate inside Pentagon networks alongside Gemini, handling both classified and unclassified documents, per PBS (PBS).

For B2B marketers and e-commerce teams selling into government, the practical signal is procurement risk: OMB guidance says systems with “severe and foreseeable risks” that can’t be adequately mitigated should be discontinued (OMB memo M-25-22). The coalition argues Grok’s record—spanning explicit image abuse, biased outputs, and prior controversial content—pushes it into that category.

If OMB intervenes, expect downstream effects: stricter vendor audits, heavier red-teaming requirements, and more scrutiny of closed-source LLM deployments in regulated and high-trust environments.