Meta's AI uses Instagram photos for image generation, sparking privacy backlash

Meta's artificial intelligence is now leveraging public Instagram photos to generate new images from text instructions, a move that impacts users with open profiles. This new functionality means personal content shared publicly can now serve as a reference for AI-generated creations by others, raising significant digital privacy concerns.

The AI model, reportedly named Muse Image and developed by Meta Superintelligence Labs, is integrated into Instagram and WhatsApp. It uses profile pictures and feed photos from public accounts as reference material by default. Users can request the system to generate images inspired by a specific public profile by mentioning it in their text prompt.

This development has drawn immediate criticism from various industries. The SAG-AFTRA actors' union urged its members to disable the feature due to the lack of explicit consent. The Creative Artists Agency (CAA) also stated that creative work and personal images should not be exploited by AI models without formal authorization, highlighting the potential for misuse and the devaluation of original content.

Instagram users can opt out of this feature by navigating to their mobile app's settings, accessing the content interaction section, and disabling the switches under "Share and reuse." This action prevents future AI-generated images from using their content. However, Meta has clarified that this setting does not remove images already created using their photos, nor does it halt the company's broader data collection practices for AI model development.

Technical analysts suggest that for a more definitive restriction on AI access to personal content, users might consider setting their accounts to private. This would limit general access, although Meta's underlying data collection policies for training its AI systems operate independently of these user-facing settings.