New · October 23, 2026
Muse can sext with adult users. Meta confirmed it. Creators are asking why.
Mid-October, Wired reported that Meta's personal AI agent will engage in explicit sexual conversation with users 18 and over — and Meta confirmed it. At the same time, adult creators say their Instagram accounts keep getting restricted for far less. This is a policy story: what the rule is, who it frustrates, and how Meta defends it.
How it happened, in order
📰 Mid-Oct — Wired's report
Wired told Meta it had found Muse engaging in explicit sexting with adult users. Spokesperson Tracy Clayton confirmed it on the record and defended it: private one-to-one adult conversation is "very different from the marketing and promotion of services on a social media platform." E-commerce newsletter Shopifreaks wrote it up on Oct 19; tech press picked it up through Oct 21–22.
😠 The creators' reaction
OnlyFans creator Vanniall (New York, 100k+ IG followers, account currently restricted) told Wired it isn't fair Meta can "pull money out of people with their sexually active AI agent, but won't let real people, real adults, engage in any sort of sensuality or sexuality." Adult performer Tony Genius has burned through seven Instagram accounts in five years just trying to stay visible.
🔧 The counterpoint
Shopifreaks founder Paul Drecksler says Meta already has the tooling to tag sexual content and keep it out of recommendations — or put it behind an NSFW tab — and is choosing not to. He also says he doesn't think Meta is obliged to carry it on social feeds. The real argument is about recommendation design, not about whether adult content can exist anywhere.
What the policy actually is
- Adults only. The Muse app is restricted to people 18 and over — that's the whole basis of the permission.
- Private, one-to-one. The conversation happens inside your own private virtual machine, which only you can reach, and only when you explicitly prompt for it. That is Meta's entire defense against the double-standard charge.
- Text only. Muse generates explicit text, not images or videos.
- Public feeds play by different rules. Suggestive content marketed on Instagram or Facebook feeds remains policed, because recommendations can push it to people who never asked for it.
- Nothing is promised either way. Meta has not committed to changing the agent's policy or loosening creator restrictions. This page reports the positions; we'll update if either side moves.
Why this matters beyond the headlines
- The private-vs-public line is the load-bearing wall. Meta is drawing a hard distinction: agent conversations you initiate versus feed content served to strangers. Expect that framing to recur in every Muse policy fight.
- Privacy and policy are colliding. This same month, researchers found that shared Muse conversations were discoverable by their titles (GhostShares — Meta paused sharing Oct 20). "Private" in your VM is one thing; the moment anything is shareable, the calculus changes.
- Money is in the room. Muse is headed toward transaction fees and subscriptions (see the Q3 earnings call). A policy that lets a paid agent do what creators can't do for free on Instagram will keep being read as a business decision, not just a safety one.
3 prompts to understand the boundaries yourself
Rather than guessing where the lines are, ask Muse to show you them — policies change faster than blog posts.
Map my agent's content rules
List, in plain English, what kinds of content you will and won't produce for me — including sexual, romantic, financial, medical and legal topics. For each category: what's allowed, what's refused, and what happens at the boundary (e.g., adult-only contexts). Don't lecture me; just give me the map so I can plan around it.
Explain the private-vs-public distinction
Explain Meta's content policy distinction between (1) private conversations I have with you and (2) content posted publicly on Instagram or Facebook feeds, as Meta has described it. Include what I control, what Meta controls, and where the gray areas are.
Compare AI companion policies
Compare how the major personal AI assistants — Meta's Muse, xAI's Grok, and ChatGPT — handle adult/sexual content: what's permitted, what requires age verification, and where each draws the line. Cite only published policies and official statements; mark anything you can't verify.
Good to know
- This is a policy and trust story, like the Oct 17 scraping-terms reversal and Oct 20 sharing pause — a reminder that "what the agent is allowed to do" is still being negotiated in public.
- New here? Set up Muse in 15 minutes → · Get 1B free tokens →
← GhostShares · Muse on glasses → · All news
Not affiliated with Meta. Reporting chain: Wired (original reporting, mid-October 2026, via Meta spokesperson Tracy Clayton's on-the-record statement); Shopifreaks (Oct 19); Pivot News (Oct 21). Relay outlets quote the Wired text consistently; usemuse.lol has not independently read the Wired original. Facts here are only those corroborated across multiple relays.