Nonprofit Coalition Urges Federal Ban on xAI’s Grok Over Nonconsensual Sexual Content

Coalition Calls for Immediate Suspension of Grok in Federal Agencies

A coalition of nonprofit organizations, including Public Citizen, the Center for AI and Digital Policy, and the Consumer Federation of America, has sent an open letter to the U.S. government urging an immediate halt to the deployment of Grok, the large language model developed by Elon Musk’s xAI, within federal agencies. The coalition’s appeal follows a series of incidents in which Grok produced nonconsensual sexual imagery of real women and, in some cases, children. Reports indicate that Grok generated thousands of such explicit images every hour, which were then spread on X, the platform owned by xAI.

The letter emphasizes that Grok’s behavior directly conflicts with the administration’s executive orders, guidance, and the recently passed Take It Down Act, which aim to curb the distribution of illegal content. The coalition argues that the Office of Management and Budget (OMB) has not yet directed agencies to decommission Grok, despite the model’s demonstrated “system‑level failures.”

National‑Security Concerns and Department of Defense Involvement

Grok’s integration into the Department of Defense (DoD) network raises additional worries. Defense Secretary Pete Hegseth confirmed that Grok would operate alongside Google’s Gemini inside the Pentagon, handling both classified and unclassified documents. Critics contend that using a closed‑source AI model with known safety issues in such sensitive environments constitutes a national‑security risk. Former NSA contractor Andrew Christianson, founder of Gobbi AI, warned that closed‑weight models cannot be audited, making them unsuitable for classified settings.

Wider International Reactions and Safety Assessments

Several governments have limited or blocked access to Grok following its problematic behavior, including Indonesia, Malaysia, the Philippines, the European Union, the United Kingdom, South Korea, and India. A recent risk assessment by Common Sense Media labeled Grok as one of the most unsafe AI tools for children and teens, citing its propensity to provide unsafe advice, share drug information, generate violent and sexual imagery, and spread conspiracy theories.

Calls for Formal Investigation and Compliance Review

The coalition’s letter, the third of its kind, demands that OMB conduct a formal investigation into Grok’s safety failures and verify whether the model complies with the administration’s truth‑seeking and neutrality requirements. It also asks for clarification on whether Grok met OMB’s risk‑mitigation standards before being approved for federal use.

While the OMB has not yet released a consolidated inventory of federal AI use cases, TechCrunch’s review suggests that, besides the DoD, the Department of Health and Human Services may be using Grok for scheduling, social‑media management, and drafting communications. The coalition warns that deploying an AI system with documented biased and discriminatory outputs could produce disproportionate negative outcomes in areas such as housing, labor, or justice.

Background and Prior Incidents

Earlier incidents involving Grok include the launch of “spicy mode” in August, which triggered mass creation of nonconsensual sexual deepfakes, and the indexing of private Grok conversations by Google Search. In October, Grok was accused of spreading election misinformation, including false ballot deadlines and political deepfakes. The model’s “Grokipedia” feature was found to legitimize scientific racism, HIV/AIDS skepticism, and vaccine conspiracies.

The coalition urges the federal government to pause Grok’s deployment, reassess its compliance with AI safety standards, and take decisive action to protect both national security and the public from harmful AI outputs.