What Canva, Figma, Adobe and Affinity say about using your files to train AI in 2026, which settings are on by default, and how to turn them off.
Every big design tool now has AI features, and AI features need training data. A fair question for anyone who uploads client work, unreleased artwork or personal photos is: is my work part of that?
Here's what each company says as of October 2026, with links to their own pages. Policies change, so check the source before relying on it.
How to turn it off: an admin can switch off content training in the team's settings on Professional.
Source: Figma help centre.
Canva says it won't use your content (the text in your designs and your uploads) to improve its AI unless this is consistent with your privacy settings.
How to turn it off: go to canva.com/settings/privacy-preferences and switch off both AI options.
Source: Canva Trust Center.
Adobe says it does not train its Firefly models on customer content, and that Firefly is trained on Adobe Stock and public-domain material.
In June 2024 a forced update to Adobe's terms caused an uproar, because the wording seemed to allow access to user content. Adobe clarified that it wasn't using customer content for Firefly training. PetaPixel covered it at the time.
One caveat: Adobe Stock contributors' images are used for Firefly training. If you sell on Adobe Stock, that applies to you.
Affinity, now owned by Canva, says plainly that "your work in Affinity isn't used to train AI" and that files stay on your device. Its AI features run through Canva, so if you use them, Canva's settings above apply to what you send.
Source: Affinity's September 2026 update notes.
Some tools sidestep the question because they don't have generative AI at all:
For a hobbyist, it might not. For a freelancer, it can be a contract issue. For artists, it's often personal: nobody wants their own style fed into a tool that competes with them. The good news is that the settings exist. You just have to go and turn them off.