Generative AI Guidelines
This set of guidelines gives you the latest information on how to meet Adobe’s requirements around Generative AI.
Introduction
Our Generative AI guidelines include:
- Acceptable content
- Restricted content
- Transparency when using Gen AI content
- Testing and security
We will edit and update this guide as our requirements and the technology evolve.
Overview
The rise of Generative AI offers potential benefits for add-ons, particularly in facilitating content creation and accelerating workflows.
However, it also means it is possible for users to be exposed to inappropriate or offensive content on a large scale, even if the add-on’s own code bundle and assets technically meet requirements.
Adobe encourages users to be able to choose whether they want to try add-ons that employ Generative AI technology. But we want to make sure that this is clearly stated upfront - and that the content they receive is appropriate and respects intellectual property (including copyrights and trademarks).
We created these guidelines to assist any developers who wish to take advantage of Generative AI in their add-ons, to make sure they do so in accordance with Adobe standards.
We have split this section into:
- Requirements (which must be followed)
- Recommendations (which are suggestions we encourage for best practice)
If your add-on does not meet required criteria, your submission will be rejected. If an issue is discovered, and your add-on is already listed in the marketplace, you’ll be contacted and given thirty (30) days to resolve the issue or risk being removed from the marketplace.
Add-ons that are listed privately are not required to undergo automated testing. However you should not use private add-ons as a way to avoid ensuring the safety of your add-on when sharing it with users.
If we receive reports of abuse, we may elect to remove an add-on whether it is private or public.
