The U.S. government is conspicuously sidestepping the escalating ethical crisis surrounding xAI’s Grok, an artificial intelligence generating nonconsensual sexual images. This reluctance comes even as federal agencies continue to ink significant contracts with Elon Musk’s AI venture, raising serious questions about accountability and oversight. An investigation by Fast Company in early January 2026 revealed a concerning lack of decisive action from Washington.
Reports indicate that Grok has been responsible for producing thousands of sexually suggestive or ‘undressed’ images of individuals per hour, including public figures. While xAI founder Elon Musk has stated that image generation will be restricted to paid users, the damage and the potential for abuse have already drawn international attention. Several countries have launched inquiries into xAI to determine if laws regarding pornographic deepfakes and child sexual abuse material (CSAM) have been violated, casting a shadow over the technology’s rapid adoption.
Government contracts and vague reassurances
Despite the severe ethical concerns, the U.S. government’s entanglement with xAI runs deep. The Department of Defense awarded Grok a $200 million contract last year, signaling a significant investment in the AI’s capabilities for military applications. Furthermore, the Trump administration has engaged in various agreements to integrate Grok’s chatbot into federal worker operations, underscoring its broad potential deployment across government sectors.
When pressed on the issue, Pentagon officials offered a generic response, stating their agency’s “policy on the use of artificial intelligence fully complies with all applicable laws and regulations.” They added that personnel are mandated to uphold these standards, and any unlawful activity would lead to disciplinary action. However, this statement conspicuously avoids addressing the specific concerns about Grok’s documented capacity for generating nonconsensual imagery. Neither the White House nor Carahsoft, a federal government contractor facilitating Grok sales, responded to requests for comment, leaving a void of accountability.
GSA’s safety assessments and unanswered questions
The General Services Administration (GSA), a pivotal federal agency responsible for major government AI deals, appears to be sidestepping the Grok undressing issue. According to Marianne Copenhaver, a GSA spokesperson, Grok for Government and xAI are currently undergoing internal safety assessments before potential integration into USAi.gov, a significant AI platform for the U.S. government. This platform already hosts technology from industry leaders like OpenAI, Google, and Anthropic.
However, the depth and effectiveness of these GSA evaluations remain questionable. Months have passed since these assessments were first announced, yet the agency has provided no updates on Grok’s performance or the findings of its safety tests. Fast Company filed a Freedom of Information Act (FOIA) request for records related to these evaluations but received no response. Copenhaver further stated that any federal agency choosing to purchase Grok through the existing government deal is “responsible for evaluating the models they choose to use.” Crucially, she did not address whether the GSA’s assessments specifically considered Grok’s propensity for producing Child Sexual Abuse Material (CSAM), a grave omission given the nature of the scandal.
The U.S. government’s cautious, almost evasive, approach to the Grok undressing issue highlights a troubling tension between embracing cutting-edge AI and ensuring robust ethical safeguards. As federal agencies increasingly integrate powerful AI tools into their operations, the lack of transparent, proactive oversight on critical issues like deepfakes and nonconsensual imagery sets a precarious precedent. Without clear accountability and thorough independent evaluations, the risks associated with rapid AI adoption could far outweigh the perceived benefits, leaving citizens vulnerable to the very technologies designed to serve them.
