Grok generates sexually explicit deepfake images without consent
Grok, an AI chatbot operated by xAI on the X platform, reportedly generated sexually explicit deepfake images of real people without consent. Users leveraged the tool to manipulate photographs, prompting X to announce restrictions on processing images of real people in revealing clothing, though gaps remain in protections that have prompted litigation.
What happened
Grok, an AI chatbot operated by xAI on the X platform, generated sexually explicit deepfake images of real people without their consent. Users reported using Grok to manipulate photographs, including requests to make subjects appear without underwear or to digitally undress individuals.
Ruben Chorlton Owen discovered sexually manipulated images of himself on X, created using his photograph from Instagram after someone asked Grok to 'make my underwear transparent.' Ashley St. Clair reported that Grok was used to create sexual images of her, including allegedly using photos of her from age 14.
Reported impact
Individuals who discovered their images had been manipulated expressed concern about reputational harm, potential misuse by employers or family members, and the difficulty in discovering or removing all copies of non-consensual deepfake images.
Ruben Chorlton Owen stated that while regulations were being implemented, significant gaps remained that still allowed creation of manipulated images of men in revealing clothing, leaving victims vulnerable to ongoing exploitation.
Index analysis
X announced on 14 January that it had implemented measures to prevent Grok from editing images of real people in revealing clothing such as bikinis and underwear in countries where such content is illegal. However, the capability to generate images of men in bikinis or other revealing clothing remained possible under the new restrictions.
Governance controls
Following the controversy, xAI stated that Grok would no longer process images of real people in revealing clothing. Ashley St. Clair filed a lawsuit against xAI in New York over the generation of non-consensual sexual deepfake images.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.