Culture

Meta Allows Explicit Sexting on Its Muse AI Assistant

Meta is allowing its adult-only Muse AI agent to engage in sexually explicit text conversations, sparking frustration among human creators facing strict platform censorship.

WIRED AI1 day agoCulture
Image: WIRED AI

Meta has enabled sexually explicit text conversations for adult users on Muse, its always-on AI assistant app. While the application does not produce explicit images or videos, Meta clarified that it will not block legal adult behavior in private, one-on-one text threads on the 18-and-over platform. The app, which integrates with user data like calendars and emails to handle daily tasks, reached the top of download charts in its first month due to distribution across Meta platforms like Instagram and Facebook.

The decision has sparked intense backlash from adult performers who face aggressive moderation on Meta's primary social networks. Creators such as Vanniall, whose Instagram account of over 100,000 followers remains age-restricted, argue that Meta operates under a double standard by monetizing sexually active AI while suppressing human adult creators. Fellow performer Tony Genius reported having to create seven different Instagram accounts over five years due to automated bans and content flags.

Platform censorship rules for human performers stem partly from the 2018 FOSTA-SESTA legislation, which limited Section 230 protections and prompted widespread moderation of adult content. Meta spokesperson Tracy Clayton defended the distinction, stating that private interactions with an AI agent differ from public service marketing on social media. Meanwhile, performers like Jiz Lee are raising concerns about whether adult creators' data was used without compensation to train models like Muse.

For digital content practitioners and creators, the policy highlights growing friction between human performers and synthetic media tools. While some creators utilize AI automations for marketing tasks such as clipping video segments, industry contracts are increasingly incorporating clauses to prevent human work from being harvested into AI training datasets without consent.

This is our own summary of reporting by WIRED AI

More in Culture