A bug has emerged in Instagram's labeling system designed to identify content created by generative AI.
Instagram AI Labeling Bug: Non-AI Images Incorrectly Flagged as Generative AI
This article is a translation. Read the Japanese original
According to The Verge, cases have been reported where the "AI Content" label is erroneously applied to images that users did not create using generative AI tools.
The causes for the label being applied are multifaceted. Instances have been confirmed where images edited using tools like Canva, which features background removal functions, are being labeled.
Canva has acknowledged that its supportive AI tools were being tagged as generative AI and explained that the issue has now been fixed.
On the other hand, there are also cases where actual AI-generated images, which should originally be labeled, are not receiving the label.
Meta stated that it performs detection by scanning metadata such as IPTC and C2PA, but specific details regarding the exact criteria and methods used for determination have not been disclosed.
--- Source: Instagram’s AI detection is a mess (again) (The Verge AI, 2026-09-04)