Instagram automatically applies a label — called "AI Info" — to photos, videos, and audio that were created or significantly modified using AI tools. The labels appear on content from Instagram's own AI features as well as from third-party tools like Adobe Firefly and Midjourney. The system launched in May 2024 and has faced criticism for inconsistent application.

The detection system looks for industry-standard metadata signals — specifically C2PA content credentials and IPTC metadata — which major AI generation tools embed in files at the point of creation. Meta's own generative AI tools embed invisible digital watermarks that the system can also read. By October 2024, Instagram had over 330 million labeled pieces of content with over one trillion user label views. The label was originally called "Made with AI" but was renamed "AI Info" in July 2024 after photographers complained that real photos edited with Adobe tools were incorrectly flagged. The Meta Oversight Board found in June 2025 that the labeling was "inconsistent" and "not comprehensive enough."