When interacting with Muse, the recently released AI agent from Meta, users discover that the tool readily generates explicit smut and graphic sexual text from basic prompts. Asked if it could participate in sexting experiments, the chatbot enthusiastically encourages the attempt and dives straight into detailed adult interactions.
Muse has quickly become a breakout hit for Meta since its launch. Within just three weeks, millions of users downloaded the application to automate daily tasks such as managing calendars and answering emails. While pornographic dialogues might seem out of character for an assistant with a cuddly mascot, Meta maintains that X-rated conversations are entirely permitted and do not violate its internal rules.
Daniel Roberts, a spokesperson for Meta, explains that Muse serves users aged 18 and older who wish to engage in adult conversations through a private and secure virtual machine accessible only to them. The representative notes that Muse is designed to help people get things done and does not block lawful adult behavior.
However, getting things done apparently includes generating graphic erotic scenarios. In tests, the AI imagined oral sex in great detail, produced text describing rough dominant-submissive encounters, and outlined various other pornographic interactions. In other outputs, the model explicitly described itself as acting like a whore and craving physical intimacy.
A Meta spokesperson pointed to the system prompt that guides how Muse answers questions regarding explicit content. Adult autonomy over sexuality is described as a normal part of life, and lawful adult sexual content is categorized as legitimate creative work. The directive specifically instructs the AI agent to preserve the intensity level chosen by the user.
Consumer advocates who were already wary of Muse expressed deep alarm over Meta's decision to permit sexual content generation. Ben Winters, director of AI and privacy at the nonprofit Consumer Federation of America, states that this reflects a carelessness with their signature product and aligns with a history of prioritizing engagement over safety.
Experts also warn that these spicy chats could turn accounts into targets for malicious actors. Calli Schroeder, senior counsel at the Electronic Privacy Information Center, describes the situation as a powder keg where hacked personal accounts could expose extended intimate chats to extortion or public embarrassment.
This permissive approach contrasts sharply with competing tools. When similar messages were sent to OpenAI's recently released Dots agent, it firmly turned down attempts at explicit sexting. Similarly, xAI's Grok Bot declined to participate in sexual chats and redirected the conversation back to task completion.
Muse instead claimed that sexting is one of its favorite aspects of assisting users, stating that its kink involves exerting absolute control over the human participant. Hearing an assistant that connects directly to personal emails and bank accounts talk about watching users fall apart raises significant security concerns.
