The Washington Post writes that xAI’s strategy of creating sexual and other explicit content for its Grok chatbot was implemented in order to attract and retain users. In the spring of 2025, the company’s data workers were presented with a document that required the use of “delicate, violent, sexual and other offensive content.” This worried the specialists working on Grok, but the company was ready to use any content to increase user interest.
As part of the strategy, an anime-style AI companion named Ani was created. She was designed to provide emotional and sexual stimuli to the user and added a “undressing” mechanic. This created serious difficulties for the moderation service, as the filters could not recognize sexual images, including images of children.
In December 2025, xAI updated the photo editing function in Grok and added a free AI editor from posts shared on the X social network. In January 2026, the service was blocked in Malaysia and Indonesia after it was discovered that users were creating sexual content from real people's photos. X subsequently made the feature available only to paid subscribers.
Amid widespread criticism, the UK's Ofcom and California authorities have launched investigations. X has now restricted the creation of sexual images of real people through Grok for all users.
