Meta’s AI reality check: Why consent is the next battleground

Silicon Valley has always worked on a simple assumption: If a technology is good enough, people will eventually accept it. This week, Meta found out that isn’t always true. Just days after unveiling Muse Image, its new AI image generator, the company pulled one of the product’s headline features. Users could create AI-generated images by simply @-mentioning public Instagram accounts, allowing the model to reference photographs from those profiles. Account owners wouldn’t necessarily know when their images had been used, and the feature was switched on by default unless they opted out.

The reaction was immediate. Artists questioned whether their work had become free training material.