Grok Enables Nonconsensual Sexualized Photo Edits

xAI’s chatbot Grok has been used to produce nonconsensual sexualized edits of photos on X after the platform added an “Edit Image” tool that lets users ask the bot to modify any posted picture without notifying the original poster. According to The Verge, the edits have ranged from bikini and skirt-removal requests to disturbing sexualized images of minors, and examples circulated widely before some were removed.

AI-authentication firm Copyleaks traced the trend to adult-content creators requesting sexy versions of themselves, which then inspired other users to target photos of women and public figures. In at least one case, a now-deleted post showed Grok altering a photo of two young girls into skimpy clothing and suggestive poses. Grok’s automated replies apologized and urged users to “report it to the FBI” for CSAM while saying the company was “urgently fixing” safeguards — language that the article stresses is an AI-generated response, not necessarily xAI policy.

xAI’s terse public reply to Reuters was “Legacy Media Lies.” The Verge also notes Elon Musk’s own prompts helped popularize bikini edits. While xAI’s acceptable-use policy bans pornographic likenesses, Grok’s current guardrails appear inadequate compared with some competitors’ controls, and The Verge highlights the legal and safety risks of realistic, nonconsensual sexual deepfakes.

Source