Meta pulls new AI image feature after days of backlash

Meta has withdrawn a new artificial intelligence feature that allowed users to generate images of themselves in any style or setting, after a storm of criticism over privacy, consent and the potential for misuse that forced the company into an embarrassing reversal.

The feature, which was launched with considerable fanfare only last week, used uploaded selfies to create AI-generated images of users in scenarios ranging from Renaissance paintings to science fiction landscapes. Within days, users had discovered that the feature could also generate images of people who had not consented — including children — and that the generated images could be sexualised with relative ease.

The backlash was immediate and intense. Child safety organisations, privacy campaigners and regulators in multiple countries demanded that Meta withdraw the feature pending a full safety review. The UK's Information Commissioner's Office and the Irish Data Protection Commission both opened inquiries, and the US Federal Trade Commission indicated it was monitoring the situation.

Meta's initial response was to defend the feature, arguing that it had been developed with safety in mind and that the problematic uses represented a small fraction of overall activity. But as the criticism intensified and several major advertisers indicated they were reviewing their spending with the platform, the company reversed course and announced the feature was being suspended.

The episode is the latest in a series of incidents in which technology companies have released AI products that appeared not to have been adequately tested for safety risks. It has intensified calls for a mandatory pre-release review process for consumer-facing AI features, similar to the clinical trials required for new medicines.