Meta’s New AI Tool Lets Strangers Put Your Face in Their Prompts — Hollywood Says Opting Out Isn’t Enough

Meta this week rolled out Muse Image, an AI image-generation tool built into the Meta AI app, Instagram and WhatsApp that lets any user generate AI images using another person’s likeness simply by tagging that person’s public Instagram handle in a prompt, with no additional consent required. The feature is opt-out by default, automatically excluding only private accounts and profiles belonging to users under 18 [1, 4]. Talent agency CAA publicly condemned the design on July 9, calling on Meta to make likeness protection the default setting rather than the exception and demanding documented consent before any third party, including an AI model, can use a person’s name, image, likeness or voice [2, 3]. SAG-AFTRA followed by formally recommending its members, and all Instagram users, manually opt out of Muse Image, calling anything short of a “clear and conspicuous opt-in” unacceptable and “a miscalculation of public sentiment” [1, 4]. Meta responded that it built the tool with “strong controls and safety guardrails from day one,” noting that adult users with public accounts can opt out in a couple of clicks [1, 3].

Why It Sucks:

Actors & Performers (SAG-AFTRA)

• Your face becomes fair game by default. Any public Instagram account, which is a professional necessity for most working actors, can be tagged into an AI-generated image with zero notification or consent required before the fact [1, 4].
• The burden falls on the performer, not Meta. SAG-AFTRA argues that requiring people to actively dig into settings and opt out inverts who should bear responsibility for protecting someone’s likeness from AI misuse [1, 4].
• Guardrails only cover minors and private accounts. Everyone else, including public figures who rely on visibility for their careers, is automatically enrolled the moment Muse Image launches [1, 3].

Talent Agencies (CAA)

• Client contracts get undercut by a platform default. CAA represents clients whose commercial value depends on controlling who profits off their image, and an opt-out system lets Meta monetize that value before any agency can intervene [2, 3].
• Consent becomes an afterthought instead of a requirement. The agency wants documented, affirmative consent before any use of a client’s name, image, likeness or voice, arguing Meta’s approach treats that standard as optional [2].
• One platform’s policy sets an industry template. If Meta’s opt-out model stands unchallenged, agencies fear competing platforms will copy the same low-friction, low-consent design rather than building in permission from day one [3].

Meta

• A flagship AI launch is already under siege. Days after rolling out Muse Image across the Meta AI app, Instagram and WhatsApp, the company is fielding coordinated public condemnation from a major union and a top talent agency [1, 2, 3].
• Built-in safeguards are getting buried by the backlash. Meta points out that private accounts and under-18 users are automatically excluded and that opting out takes only a couple of clicks, but that nuance is getting lost in the “non-consensual AI” framing [1, 3].
• Pressure to flip the default could gut the feature’s design. Muse Image’s viral, tag-anyone mechanic is central to driving adoption, and forcing an opt-in model, as CAA and SAG-AFTRA are demanding, would sharply limit how many faces are actually available to generate with [2, 4].

Sources & Citations:

[1] Deadline: SAG-AFTRA Recommends Members Opt-Out Of Meta’s AI Feature: “Take Action To Protect Your Likeness”
[2] Variety: CAA Slams Meta for Using Opt-Out Policy for AI Platform Muse Image
[3] Deadline: CAA Calls BS On Meta’s AI Image Generator; Zuckerberg Pushes Back On Muse Privacy Worries
[4] Variety: SAG-AFTRA Slams Meta AI for Using Your Instagram Photos, Says Opt Out