← Latest briefing

Technology

Instagram users report inaccuracies with automated AI content labels

Meta's detection system is reportedly labeling standard photos as AI while missing some synthetic media.

The short version

  • Instagram users report that Meta's automated labeling system is erroneously tagging non-AI images with "AI Content" badges.
  • Unlabeled synthetic imagery is reportedly bypassing detection and remaining unmarked on the platform.
  • Minor digital adjustments, including using Canva's background removal feature, have triggered the automated labels.

Key facts

  • Instagram users report that Meta is automatically adding "AI Content" labels to images that were not generated or edited using generative AI tools.[The Verge]
  • Some fully AI-generated images are reportedly bypassing Meta's detection system without receiving any label.[The Verge]
  • Users reported that basic photo editing functions, such as Canva's Background Remover, have triggered the incorrect AI tags.[The Verge]

What remains uncertain

  • The full scope of the labeling errors and whether Meta plans technical adjustments to its detection system remain unclear.[The Verge]

Sources

Outlet counts describe coverage, not independent confirmation. Reports may share a wire service or original source.