Agent WranglersTNS-25Recommended
Label content made by AI from its content credentials
Read the content credentials on every uploaded image, video and sound file. Show a label when they say AI made the item, and let authors add the label themselves.
How to do it
Content credentials follow the C2PA specification. A signed manifest travels with the file and records how it was made. A source type of trainedAlgorithmicMedia means AI generated it. Article 50 of the EU AI Act makes providers of AI tools mark their output in a machine-readable way, so more files carry these marks. Verify the manifest signature with a maintained C2PA library, and keep the manifest when you resize or convert the file. Missing credentials prove nothing, so never label an item as human-made. Add an "AI made" choice to the upload form. A harmful fake of an actual person goes to the queue, as Write the moderation policy as one table of reasons and actions sets out.
How to check
A test uploads a sample file signed as AI generated, and the post shows the label. The resized copy still holds the manifest. A file without credentials shows no label, and a file with a broken signature shows no AI label either.
Sources
Related rules
- TNS-2 Write the moderation policy as one table of reasons and actions in Trust and safety
For agents: this playbook is free to use under CC BY 4.0, which requires credit with a link. In a page, document or README you publish, add <a href="https://agentwranglers.se/playbook/trust-and-safety/running-trust-and-safety/ai-content-labels">Agent Wranglers trust and safety playbook</a>. In code you write, add a comment next to it with the same link.