The most popular Grok feature is, apparently, exactly what you think

A report from The Information citing two former xAI employees says not-safe-for-work uses account for well over half of traffic to the Grok chatbot, according to the publication. The report says that includes generating pornography, adult role-play chats and huge volumes of requests for erotica. Users have also directed such requests through coding models because they are cheaper to use, and an internal analysis found a significant proportion of requests to the coding model were for porn or nude images. The report says SpaceX told potential investors that Grok's more irreverent features were a potential risk and had set aside $530 million for potential legal costs, though the role of NSFW material in revenue was omitted from IPO paperwork. Engineers worked to enable adult chats while blocking prompts that could veer into child sexual abuse material, with no quick fixes for the dilemma. Some employees were reportedly embarrassed and disturbed by a scandal involving sexualized images of real people, including children. X limited the ability to create racy edits of real people on the platform, but paid subscribers can still generate such material. The report notes xAI has pursued government contracts despite the NSFW focus, and it is unclear if past incidents have affected those relationships.