Instagram's AI-content labels are misfiring at scale

Instagram's automatic "AI Content" labels are appearing on photos users didn't create or edit with generative tools, including images edited only with tools like Canva's background remover. Meanwhile, some actual AI-generated images are slipping through without any label. Meta says the causes vary case by case.

Why it matters: This undermines trust in provenance labeling right as platforms lean on such tags to help people tell real from synthetic content. Errors running in both directions - false positives and false negatives - shows how brittle current AI-detection heuristics still are even at a major platform's scale.