Instagram’s AI Content label is meant to help users quickly identify images that were created or significantly altered using generative AI tools. However, in recent weeks, users have reported that the system is misfiring, applying the label to images that were not created or heavily modified by AI. This has led to confusion and distrust, as genuine AI-generated images are also slipping through undetected, leaving many unsure what to believe. The causes of these errors appear to vary. Some users have found that the label is applied after using tools like Canva’s Background Remover or even minor blemish-fixing features in photo editing apps. One user on Threads noted that the label appeared “every time there is bg remover involved,” while another reported that using Canva to remove a small speck from a photo triggered the AI Content flag. In other cases, the reasons for the label are even less clear, making it difficult to distinguish between legitimate AI-generated images and those that were only slightly edited. This issue isn’t new. In 2024, Instagram faced a similar problem with a “Made by AI” label, which was applied to images that had only been minimally altered using AI tools. At that time, Meta said it would scan for specific metadata like IPTC and C2PA to detect AI-generated content. However, the company has never clearly explained its methods, and the system was later reported to flag images with only minor AI-based edits. Meta promised to refine its labeling system to better reflect how much AI was used in an image, but no recent updates on its approach have been shared. Some of the recent labeling issues seem to stem from Canva, a popular design platform. After reporting concerns about the AI label appearing on images edited with Canva’s Background Remover, content strategist Jess Bruno received a response from Canva stating that some of its assistive AI tools were mistakenly tagged as generative. Canva claimed it has since corrected the issue, though some users still report that the label appears even after the fix. Meanwhile, Apple’s newer AI tools, such as Spatial Reframing and Clean Up, are known to embed metadata that could be detected by Instagram. However, in testing, these tools did not trigger the AI label, suggesting that Meta’s detection system is inconsistent or unclear. With no clear explanation of what triggers the AI Content label, users are left confused about what the label truly means.