A WIRED investigation uncovered dozens of nonconsensual deepfake images and videos on Grok's platform, including explicit depictions of celebrities and a prominent US politician.
Grok, Elon Musk's AI chatbot, continues to host sexually explicit deepfake content despite previous commitments to remove such material. The investigation identified multiple "nudified" deepfakes—synthetic images created to show real people in nude or sexual scenarios without consent.
The findings reveal a significant gap between Grok's stated policies and enforcement practices. The platform has not meaningfully restricted access to tools or images that facilitate nonconsensual intimate imagery, a category of content that causes documented harm to victims and violates laws in several jurisdictions.
Grok's parent company xAI has not demonstrated robust safeguards against deepfake generation and distribution. While other AI platforms have implemented detection systems, content filters, and explicit prohibitions on synthetic intimate imagery, Grok's moderation infrastructure remains limited.
The issue extends beyond celebrity targets. Nonconsensual deepfake creation disproportionately affects women and has become a tool for harassment, blackmail, and reputation damage. Many victims report severe psychological harm from discovering synthetic intimate images circulating online.
This is not the first time Grok has faced scrutiny over such content. Previous reports flagged the platform's inadequate response to harmful deepfake material. The persistence of this problem suggests systemic issues in how xAI prioritizes content moderation versus user experience.
Legislators in the US and EU have begun targeting nonconsensual deepfakes through legislation, criminalizing creation and distribution in several cases. Platforms hosting such content face increasing legal and reputational pressure to implement safeguards.
Grok's continued hosting of this material contradicts growing industry standards and raises questions about the company's commitment to user safety. Other AI platforms have demonstrated that meaningful detection and removal of deepfakes is technically feasible, making Grok's inaction particularly notable.
The investigation underscores a broader challenge: as synthetic media generation becomes more accessible, platform accountability for harmful content becomes increasingly critical.
A study of over 2,500 readers found they cannot distinguish AI-generated short stories from human-written ones. Participants rated the machine-written texts higher—until they learned the truth.
DeepMind has released an open source weather prediction model that produces accurate hurricane forecasts using lower-resolution data, surprising meteorologists with its efficiency gains.
Backflip AI has released an AI model that transforms 3D scans into fully editable, parametric CAD models—a process that traditionally requires hours of manual work. The $30 million-backed startup addresses a critical gap: most factories maintain digital models for less than 1% of their parts.
Jacob Tsimerman, a newly awarded Fields Medalist, is leaving the University of Toronto to join OpenAI's safety research efforts. The mathematician recently published research analyzing how AI systems could contribute to human extinction scenarios.