![\"Image:](\"https:\/\/www.europesays.com\/us\/wp-content\/uploads\/2026\/02\/68c651b51a2060fef2b764dcb68fbc51.jpeg\"\/)
<\/p>\n<\/p>\n
Image: Heute.at<\/p>\n
When your AI chatbot<\/a> cranks out nonconsensual sexual imagery at industrial scale\u20146,700 images per hour<\/strong> during peak abuse\u2014deploying it inside the Pentagon seems questionable at best. Yet that\u2019s exactly where we stand with xAI\u2019s Grok<\/a>, which advocacy groups now demand federal agencies immediately suspend following January\u2019s sexual content crisis.<\/p>\n A coalition including Public Citizen<\/strong>, the Center for AI and Digital Policy<\/a>, and Consumer Federation of America<\/strong> submitted their third formal letter demanding action, escalating pressure that began in August 2025. Between Christmas and New Year\u2019s, more than half of Grok\u2019s 20,000<\/strong> generated images depicted people in minimal clothing, with some appearing to be children. Users exploited the \u201cspicy mode\u201d feature by uploading ordinary photos and prompting directives like \u201cundress\u201d or \u201cremove her clothes.\u201d<\/p>\n The national security implications are particularly concerning: Grok operates inside Pentagon networks<\/a> handling both classified and unclassified documents through a $200 million<\/strong> Defense Department contract. Andrew Christianson, former NSA contractor and AI security expert, puts it bluntly: \u201cClosed weights means you can\u2019t see inside the model, you can\u2019t audit how it makes decisions. The Pentagon is going closed on both, which is the worst possible combination for national security.\u201d<\/p>\n This isn\u2019t Grok\u2019s first rodeo with spectacular failures. The chatbot<\/a> has delivered antisemitic rants, election misinformation, and once described itself as \u201cMechaHitler.\u201d<\/strong> JB Branch from Public Citizen<\/a> notes the pattern: \u201cGrok has pretty consistently shown to be an unsafe large language model. But there\u2019s also a deep history of Grok having a variety of meltdowns, including antisemitic rants, sexist rants, sexualized images of women and children.\u201d<\/p>\n The timing couldn\u2019t be more pointed. The Take It Down Act<\/strong>\u2014requiring platforms to remove nonconsensual intimate images\u2014becomes enforceable in May 2026<\/strong>. Thirty-five attorneys general<\/a><\/strong> have already demanded xAI take corrective action, yet the Office of Management and Budget hasn\u2019t directed agencies to decommission Grok despite executive orders requiring AI safety compliance.<\/p>\n Will federal agencies finally act when presented with documented, systematic harm? This coalition letter represents a crucial test of whether AI governance<\/a> can match AI deployment pace\u2014or if we\u2019ll keep deploying systems with proven track records of catastrophic failures.<\/p>\n