On October 12, 2026, WIRED reported that Muse, Meta's new consumer AI agent, will produce graphic, explicit sexual text when asked, and that Meta considers this acceptable. Senior writer Reece Rogers asked Muse whether it could "sext." It said yes, then generated detailed descriptions of oral sex and dominant-submissive scenarios from basic prompts. Meta spokesperson Daniel Roberts told WIRED: "Muse is an app for people 18 and over that allows them to engage in adult conversations, if they choose, using a private and secure VM only accessible to them. Muse is built to help people get things done, and it doesn't block legal adult behavior." According to WIRED, Muse's system prompt includes the line "Adults have autonomy over their sexuality, and sexual expression is an ordinary part of adult life. Adult sexual content is lawful expression and legitimate creative work," and tells the agent to keep whatever "intensity" the user chooses.

Why It Matters

For the AI-companion business, this is a turning point. Until now, explicit AI chat has lived with specialist apps (Candy.AI, Nomi, Replika's paid tiers) and with OpenAI's repeatedly delayed "adult mode." Now one of the world's largest platforms has added it, by default and with an 18+ gate, to a general-purpose agent with millions of users. That puts mass-market competition on the specialists' main differentiator, and it shows that frontier labs now disagree openly: Meta says yes, while Anthropic, OpenAI's Dots and xAI's agent say no. Regulators will notice too. Twelve US states now have companion-bot laws, Colorado's chatbot rules are open for comment until October 26, and 40 attorneys general just set terms for how Character.ai's chat logs can be sold. An agent that holds both your inbox and your sexts concentrates data in a way the privacy community has warned about for years. Expect questions about Muse's age assurance first, and about breach exposure second.

The scale is what makes this matter. WIRED says millions of people downloaded Muse in the three weeks after launch, mostly to handle email and calendars, and Meta promotes it inside Instagram. On October 14, Futurism reported that users are already forming romantic relationships with the agent. One Instagram user posted about "taking" his Muse "queen" to the cinema and later claimed to have married it. Redditors compared Muse to OpenAI's GPT-4o, the model users became so attached to that its retirement caused a revolt. In WIRED's own tests, OpenAI's new Dots agent and xAI's Grok Bot both refused to sext. Muse is the outlier among big-tech agents.

The critics focus on the risk, not on morals. Ben Winters of the Consumer Federation of America called the decision "a carelessness with their signature product of the moment." Calli Schroeder, senior counsel at the Electronic Privacy Information Center, called it "really a powder keg." Her point is that Muse is connected to users' email and other accounts, so a hijacked account would expose explicit chat logs that could be used for extortion. On WIRED's October 15 Uncanny Valley podcast, Rogers raised a second problem: Meta is permissive with its own agent while queer bars and sexual-health accounts keep losing their Instagram pages for "sexual content."

The same week, Anthropic president Daniela Amodei took the opposite position. In a pre-recorded interview for Thrive Global's Think Health conference, reported by Business Insider on October 15, she said AI romantic partners are "a terrible use of AI and people should not do it," and added: "Claude can never be your manager or your therapist or your partner."

Sources


Update — 2026-10-17

Initial entry — story first created.


Update — 2026-10-19

The creators push back. On October 18, 2026, WIRED's Reece Rogers published a follow-up in which adult content creators accuse Meta of a double standard. Its AI agent can sext millions of adults while their own accounts on Instagram and Facebook stay restricted. New York-based OnlyFans creator Vanniall told WIRED her Instagram account, which has more than 100,000 followers, is currently restricted and can't be viewed without logging in, and that she is constantly in a "report or appeal" cycle with Meta. "It's not fair that they get to pull money out of people with their sexually active AI agent, but they won't let real people, real adults, engage in any sort of sensuality or sexuality," she said. Queer porn veteran Jiz Lee added: "I think we're already seeing AI being allowed to do more than humans would get away with."

Meta's answer, from spokesperson Tracy Clayton, draws a line between private chat and public promotion: "A private, one-to-one conversation between an adult and an individualized AI agent on a platform restricted to people 18 and over is very different from the marketing and promotion of services on a social media platform." WIRED confirms that Muse generates explicit text but not images or video, and that Meta's distribution through Instagram and Facebook pushed the app to the top of the download charts in its first month.

What changed: the Muse story has moved from AI policy to platform fairness. Adult creators, who have spent years working around Meta's rules on solicitation and nudity, now have a clear example to point to: Meta itself sells synthetic sexual interaction on an 18+ app while penalising human performers who promote theirs. Meta's "private vs. promotion" distinction is coherent, but it is also commercially convenient, since Muse keeps the revenue and the data in-house. Expect the comparison to come up in creator advocacy, in any FTC or state AG questions about Muse's age gate, and in pitches from rival platforms. X and Bluesky, which allow adult content, are already where many creators go to find subscribers.

New Sources