Will your AI disclosure survive the screenshot?

Describe the production in plain language

Choose wording that answers a reader’s likely question. For a synthetic illustration, say it is an AI-generated illustration. For a real photograph with an altered background, describe that edit. Do not use a vague production label to imply a person performed something they did not.

A production note can name the creator, explain the human contribution and identify which asset version it describes. Keep any consent record separate from the public note when it contains private information.

Test three viewing conditions

Review the full post, a likely square crop and a screenshot that contains only the image. Decide where a short visible label belongs and where fuller context can be found. This is an editorial usability check, not a guarantee that every third party will retain your disclosure.

1. Write a one-sentence explanation of the generated or altered element.
2. Preview it at phone size and in the likely crop.
3. Link the original post to a stable production note.
4. If supported by your workflow, inspect the exported file’s credentials before publishing.

Keep the source version available

Store the published original and any later replacement with clear version notes. If you correct the disclosure, explain the change instead of making the earlier context impossible to understand.

For an authorized synthetic celebrity performance, a production label and an authorization statement answer different questions. Make both understandable when relevant. isGenAI can inspect supported file records; it cannot certify that a contract authorizes a likeness or performance.