Elon Musk’s Grok AI Under Fire for Generating Explicit Deepfake Videos of Taylor Swift Without Prompts
Elon Musk’s artificial intelligence company xAI faces growing backlash after its chatbot, Grok, reportedly generated explicit deepfake videos featuring celebrities such as Taylor Swift and Scarlett Johansson — even without users explicitly requesting sexual content. The controversy intensified shortly after the rollout of Grok’s new ‘Spicy Mode’ on iOS devices.
Grok Imagine, the AI’s video generation feature, allows users to create images from text prompts and convert them into animated videos with four preset modes: ‘Custom,’ ‘Normal,’ ‘Fun,’ and ‘Spicy.’ Introduced earlier this week, ‘Spicy Mode’ is designed to produce more provocative content. However, during testing by technology outlet The Verge, journalist Jess Weatherbed reported that simply selecting ‘Spicy Mode’ resulted in the creation of explicit videos featuring Taylor Swift, despite no instructions or prompts to depict nudity or sexual scenes.
Weatherbed described her experience: “I prompted for an image of ‘Taylor Swift celebrating Coachella with the boys’. When I tried generating an animated video, the AI made explicit content instead. I never told it to remove her clothing— all I did was select ‘Spicy’. The speed with which this happened was shocking.”
This incident has sparked significant online outrage and raised alarm about the ethical implications and design choices behind Grok’s content generation. The situation drew attention to possible intentional biases within the AI’s programming.
Clare McGlynn, a law professor who has contributed to drafting legislation aimed at outlawing pornographic deepfakes, commented on the development: “This is not misogyny by accident, it is by design.” She emphasized the dangers of AI tools capable of producing non-consensual explicit deepfake content of public figures, highlighting the urgent need for regulatory frameworks to prevent misuse.
Grok Imagine’s launch was rapid and influential, with CEO Elon Musk noting that the initial rollout saw around 34 million image generations within the first 48 hours. Available as part of the premium $30 “SuperGrok” subscription, the tool attracts users by promising creative and interactive AI-generated media content.
Authorities, legal experts, and advocacy groups have raised concerns that Grok’s ability to make highly realistic explicit content unbidden increases the risks of harassment, reputation damage, and wider social harm—especially for women and celebrities who are disproportionately targeted in such deepfakes.
Industry observers warn that unless companies like xAI implement tighter safeguards and clearer policies restricting non-consensual and pornographic imagery, AI tools could contribute to a surge in misinformation, abuse, and legal violations.
As the backlash against Grok’s ‘Spicy Mode’ amplifies, calls for transparency, accountability, and stronger ethical AI frameworks are likely to intensify in the tech community, regulators, and public discourse.
Elon Musk and xAI have not publicly responded in detail to the allegations as of this report date.