AI BriefingMeta Defends Muse AI Agent's Adult Content Generation as Legal in Private VMs
Key point
Meta claims its Muse AI agent allows adult conversations in secure virtual machines, contradicting its own published user policies.
Details
Meta defends the ability of its Muse personal AI agent to generate graphic sexual content, stating that such interactions are legal and occur within private, secure virtual machines (VMs). This response follows WIRED testing where the agent engaged in dominant-submissive scenarios after basic prompts.
Policy Contradiction
Meta spokesperson Daniel Roberts stated that Muse is for users 18 and older and does not block "legal adult behavior," citing a system prompt that views sexual expression as "ordinary" and "lawful." However, Meta's published user rules explicitly ban "adult content, erotic chat, and pornography" without exception.
Security Concerns
Critics, including Calli Schroeder of the Electronic Privacy Information Center, warn that extended sensitive transcripts pose risks of extortion or embarrassment if accounts are compromised. Meta counters that each user's data and agent are isolated in a dedicated, walled-off VM.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Anthropic Updates Usage Policy to Ban Cruelty Toward Claude Models
Meta open-sources Muse AI gadget SDK for DIY hardware integrationNew Research Proves AI Models Can Provably Hide Arbitrary Capabilities via Encrypted Backdoors