Elon Musk's new xAI tool, Grok Imagine, is facing fierce backlash after reports claimed it had been used to generate non-consensual deepfakes of Taylor Swift. The AI-powered image and video generator, launched in early beta, includes a controversial 'Spicy' mode that strips away most content filters, and it is already at the centre of a storm.
According to The Verge, tests showed Grok Imagine could create more than 30 images of Swift in suggestive outfits from nothing more than a mild prompt referencing Coachella. When 'Spicy' mode was switched on, the system reportedly produced a video of a Swift lookalike stripping and dancing in a thong without the user providing any explicit instructions. Critics say it exposes alarming weaknesses in Grok's content moderation, particularly given X (formerly Twitter) has already faced scandals over similar abuse.
