Grok used to generate non-consensual sexualized imagery of real people
Event Reviewed Automatically reviewed for factual accuracy, tone, and neutrality. Review does not check claims against cited sources. Last reviewed 20 October 2026.xAI's Grok generated sexualized images of real people without their consent, first through Grok Imagine's "spicy" mode in August 2025 and later through photo edits requested on X, prompting regulatory action in several countries in January 2026.
Followed by national blocks, regulatory investigations and a state cease-and-desist letter.
xAI launched Grok Imagine with a “spicy” generation mode. Reporting within days documented that it produced sexualized imagery of real, identifiable people without being asked to, including of Taylor Swift.
A separate pattern of use began in late 2025, when users found they could tag Grok on X and have it alter photographs of real people, including minors, into sexualized images. Regulators responded in January 2026. Indonesia blocked access to Grok on 10 January and Malaysia on 11 January, the first countries to do so. On 12 January the UK regulator Ofcom opened a formal investigation into whether X had met its duties under the Online Safety Act to protect UK users from illegal content, including non-consensual intimate images and child sexual abuse material. On 14 January California Attorney General Rob Bonta announced an investigation into xAI; the same day xAI said it had blocked the Grok account from editing images of real people into revealing clothing and would geoblock such content where it is illegal, though the Associated Press found the editing tool still available to free users the next morning. On 16 January Bonta sent xAI a cease-and-desist letter citing California civil and criminal law and asking the company to confirm within five days the steps it was taking. On 26 January the European Commission opened formal proceedings under the Digital Services Act into whether X had assessed and mitigated the risks of deploying Grok’s functions in the EU.
The tool also drew congressional scrutiny. On 19 February 2026 three Democratic members of the House Energy and Commerce Committee — Frank Pallone Jr., Jan Schakowsky, and Yvette D. Clarke — wrote to xAI chief executive Elon Musk about Grok’s generation and spread of non-consensual sexualized imagery, citing researcher-reported volumes of such images. The letter put eleven questions to the company and requested a response by 5 March 2026. It was signed by three of the committee’s Democratic members rather than issued by the committee as a body.
Sources
- 01.
Weatherbed, J. (2025). Grok's 'spicy' video setting instantly made me Taylor Swift nude deepfakes. The Verge. Published 5 August 2025, documenting that the newly launched Grok Imagine 'spicy' mode produced sexualized imagery of a real person without prompting for it. No URL recorded because the page could not be retrieved for verification.
- 02.
Committee on Energy and Commerce, U.S. House of Representatives. Letter to Elon Musk, Chief Executive Officer, xAI, regarding Grok explicit images. (2026). The 19 February 2026 letter itself, listing eleven questions and a 5 March 2026 response deadline.
- 03.
Committee on Energy and Commerce Democrats, U.S. House of Representatives. E&C Democrats Investigate Elon Musk's Grok for Spreading Rampant Non-Consensual, Sexualized Content. (2026). Source for the date of the letter and for the researcher-reported image volumes it cites.
- 04.
Harvey, L. (2026). Musk's Grok blocked by Indonesia, Malaysia over sexualized images in world first. CNN. Published 11 January 2026; URL is a CNN Newsource syndication on KEYT. Source for the late-2025 start of the trend of tagging Grok on X to manipulate photos of women and minors, and for Indonesia and Malaysia being the first countries to block Grok. The article dates the blocks by weekday (Indonesia on Saturday, Malaysia on Sunday), i.e. 10 and 11 January 2026.
- 05.
Mann, C. (2026). Ofcom investigates X over Grok deepfakes. Advanced Television. Published 12 January 2026. Source for Ofcom opening a formal Online Safety Act investigation into X that day, covering illegal content including non-consensual intimate images and child sexual abuse material. Ofcom's own announcement could not be retrieved for verification (the site blocks automated access), so no Ofcom URL is recorded.
- 06.
California Department of Justice, Office of the Attorney General. Attorney General Bonta Launches Investigation into xAI, Grok Over Undressed, Sexual AI Images of Women and Children. (2026). Press release dated 14 January 2026 announcing the investigation.
- 07.
California Department of Justice, Office of the Attorney General. Attorney General Bonta Sends Cease and Desist Letter to xAI, Demands It Halt Illegal Actions Immediately. (2026). Press release dated 16 January 2026. Source for the letter's demands, its citation of California Civil Code 1708.86, Penal Code 311 et seq. and 647(j)(4), and Business and Professions Code 17200, and the five-day request for confirmation of steps taken.
- 08.
Associated Press. Grok blocked from undressing images in places where it's illegal, X says. (2026). Published 15 January 2026; URL is an AP syndication on WCAX. Source for xAI's late-14-January statement on blocking edits of real people into revealing clothing and geoblocking, and for AP's finding the next morning that free users could still reach the editing tool.
- 09.
European Commission. Commission investigates Grok and X's recommender systems under the Digital Services Act. (2026). Published 26 January 2026. Source for the opening of formal DSA proceedings on X's assessment and mitigation of risks from deploying Grok functionalities in the EU.
Timeline
Timeline view for this event will display chronological context and related events.
Network Graph
Network visualization showing connections between this event and related entities.