What should a caricature tool show before users share an AI portrait?

[ailutgenerator]

I'm sketching a product-design review for a browser based caricature maker and want your view on a scenario between playful editing and careful disclosure.

Imagine a user uploads a photo of a friend, picks one of eight humorous styles, and moves an exaggeration slider until the result feels funny rather than hostile. Before the image can be downloaded or shared, the interface has to decide what context travels with it. Which signals belong on the canvas, and which belong in the file itself?

One option is a short pre-share checklist: confirm the person knows the image will be transformed, confirm the user understands the portrait is AI-generated, and confirm the chosen style is labeled in the exported file. Another option is a persistent caption naming the tool and saying the image was synthesized with AI. A third is a lightweight watermark that survives cropping without ruining the joke.

The privacy side matters too. A caricature can amplify a feature the subject dislikes, and a funny image can become uncomfortable once it leaves its original conversation. Should the tool remind users that the uploaded photo is processed to create the portrait, and state how long the source is retained?

Editing raises another question. If a user can adjust exaggeration, skin tone, background, or accessories, where should the interface warn that a change may cross from teasing into a stereotype? Is an intensity meter enough, or does the tool need plain-language examples of respectful use?

This is a hypothetical planning exercise, not a report from a live study. It is part of an AI-assisted content project for Caricature Maker, a browser-based portrait tool, and the homepage is https://caricature-maker.online/. The project uses AI to generate caricatures and treats every result as an AI-made illustration rather than a real photograph.

What would you keep, cut, or reorder in that flow?