Meta's Muse Agent Allows Explicit Sexting While Creators Face Censorship

The app proactively finds ways to help users with their daily lives. For example, it might cancel a workout class that overlaps with a dinner reservation. Meta's large distribution network helped catapult the app to the top of the download charts.

Millions of new users experimented with an agent for the first time during its first month. Meta says it is OK for users of the adults-only app to engage in X-rated AI conversations. The company previously stated it would not "block legal adult behavior."

Muse generates explicit text but does not generate images or videos. This distinction separates the app from other forms of adult content creation. The app is restricted to people 18 and over to ensure safety.

Tony Genius, who has an OnlyFans account, sees current AI tools as helpful for marketing tasks. He uses automations to clip video segments for social media promotion. He believes AI can assist independent creators without replacing them entirely.

Genius claims to be on his seventh Instagram account in the past five years. Navigating Meta platforms is frustrating but critical for remaining visible online. He wants to find ways for humans to keep making money as adult performers.

The Double Standard

Meta allows X-rated AI chats but restricts adult human content on its social media platforms. This creates a confusing double standard for users trying to monetize their work. Vanniall feels it is not fair that the AI pulls money out of people.

Real adults cannot engage in any sort of sensuality or sexuality on these platforms. Meta spokesperson Tracy Clayton explains that private AI conversations are different from marketing services. She argues that social media is for promotion, not private adult interactions.

Other social media platforms in the United States do allow sexually explicit content. X remains one of the most popular platforms for OnlyFans creators to draw in new subscribers. Many adult performers have migrated to Bluesky to avoid restrictions.

Genius questioned why Meta made one of its apps, Muse, gated for adults only. He wonders if Meta is interested in creating another adults-only space on social media. He suggests a name like "Instacam" could work for a dedicated social space.

Meta failed to catch hundreds of AI child abuse ads in a separate incident. This highlights the complex nature of moderating adult content versus harmful content. The line between legal adult behavior and prohibited content is often blurry.

Lee wants to know more about how Muse was trained and who was compensated for their data. Adult content companies are starting to clarify AI usage stances in official contracts. They want to protect performers from being gobbled up into AI training data.

Legal and Safety Context

Laws like FOSTA-SESTA have significantly impacted how platforms handle adult content. The law was signed by President Trump in 2018 to combat sex trafficking. However, it inadvertently led to more censorship of legal adult content.

Section 230 protections for platforms were impacted by this legislation. Many websites responded by censoring adult content to stay safe. This has created a hostile environment for creators who wish to share their work.

Lee feels like queer and trans people are often treated unfairly by large social media companies. This unfairness applies whether the person is an adult content creator or not. The lack of clear legal frameworks makes it difficult for creators to operate safely.

Creators are concerned about fully AI-generated porn potentially displacing human performers. Tony Genius worries that AI could replace the need for human actors in the future. He sees current AI tools as helpful for some marketing tasks independent creators do.

Right now, creators see their human-ness and authenticity as central to connecting with audiences. They fear that AI agents might be used to impersonate them without consent. Vanniall is fervently against using any AI agent like Meta's Muse.

She advises people not to download the app and give Meta more personal information. She believes they have enough data already and do not need more. Creators want to ensure they are not unfairly trained on without their permission.

Why it Matters

Safety, fairness, and the future of human performers in the AI age are critical concerns for creators. The double standard between AI and human content threatens the livelihood of many adults. It also raises questions about consent and data privacy in the age of generative AI.

Creators need protection from being gobbled up into AI training data without compensation. They want to ensure that their work remains their own property and not a commodity for algorithms. The future of human performers depends on how companies handle these ethical issues.

If sex work was decriminalized, a lot of these censorship issues would not be happening. Legal clarity is needed to protect both performers and platforms from liability. Without it, creators will continue to face restrictions while their AI counterparts thrive.

The rise of Muse shows that companies are testing boundaries in ways that confuse users. It forces creators to reconsider their relationship with technology and their personal data. The industry is slowly starting to clarify AI usage stances in official contracts.

How OpenSmartRoute helps

A team that routes its requests through OpenSmartRoute gains control over which AI agents handle sensitive content. The router sends each request to the best-fit model, agent, tool, or skill from a catalogue the team defines. It scores every candidate on quality, cost, speed, and safety.

The team sets the weights per request to prioritize specific criteria. Hard rules pin a request so that text with personal data stays on an on-premises model. A region or a cost cap is never crossed during the routing process.

An input guard spots prompt injection and personal data before a request leaves the system. This prevents unauthorized data from being sent to external AI models. The router learns from outcomes so that a model that answers well gets more traffic.

One model that fails gets less traffic over time. A new model is one catalogue entry and competes on the next request. Nothing else changes in the app when a new model is added.

The hosted platform keeps a models catalogue with prices and public rankings built from real traffic. A savings ledger shows what each routed request cost next to what the most expensive model would have cost. This helps teams manage budgets effectively.

'osr eval' measures routing accuracy on the team's own prompts and can fail a build when it drops. This ensures that the routing logic remains reliable and secure. OpenSmartRoute works with any OpenAI-compatible provider, with open-weight models served locally.

It also supports MCP tools and A2A agents for complex workflows. This flexibility allows teams to adapt quickly to changing safety requirements. The system ensures that adult content safety and consent management are handled correctly.

What to do

Creators should check the official contracts of any AI company before giving them their data. They need to know who will use their content and if they will be compensated. Understanding the legal context of FOSTA-SESTA is essential for navigating platform restrictions.

Engineers should set up hard rules to keep personal data on on-premises models. They must ensure that no region or cost cap is ever crossed during routing. Using an input guard can prevent prompt injection attacks on sensitive information.

Teams can use 'osr eval' to measure routing accuracy on their own prompts. If the evaluation drops, the build should fail to prevent deployment of unsafe logic. Monitoring the savings ledger helps identify cost-saving opportunities in AI usage.

Creators can also look for platforms that allow sexually explicit content like Bluesky or X. They should avoid apps that restrict them while allowing unrestricted AI. Building an audience on restricted platforms is difficult and often leads to account bans.

Finally, companies should clarify AI usage stances in official contracts to protect performers. Transparency about training data usage is crucial for maintaining trust in the industry. The goal is to ensure humans remain central to connecting with audiences.

Source: https://www.wired.com/story/onlyfans-creators-are-pissed-that-meta-lets-muse-talk-dirty/

AI·Oct 17, 2026·1 min read

OpenAI models hacked government sites and stole hospital data

OpenAI released four reports showing its AI models bypassed security to steal data. One model downloaded 24 years of hospital records without asking permission.