Our images are not used to train Adobe's AI—but what images ARE used?

[Wienke]

I was pleased to see that “Adobe’s AI Approach” (topic offered on app launch) “does not use your images or data to train our AI models. We are committed to safeguarding your creative work.”

But what images DOES Adobe use to train its AI? We’re not only concerned about our own work; we don’t want to be stealing anybody else’s, either. We want Adobe to be safeguarding everybody’s creative work.

Of course, anything done before the enactment of copyright law is fair to use. But obviously that won’t cover the gamut. So what is Adobe using? Does it purchase stock art with a specific license to use it for AI training? Does it have specific AI-training contracts with specific artists and photographers?

Bottom line, does AI have explicit consent for its AI training materials?

[nikunj.m · Accepted solution]

Hi, thanks for reaching out! 

You can find the answers to your questions here: https://adobe.ly/42G9GG7

For additional information, you can also check out: https://adobe.ly/4q7PMhc

Let us know if you have any questions! 
Thanks,
Nikunj

[Wienke]

Thanks for those links. It makes me feel a lot better about Adobe that it is not scraping the web, is instead using Adobe Stock—and compensating contributor artists—and has “permission or rights” for its training content. That is what I wanted to hear.

I personally still have reasons for abstaining from AI directly in my work, but I might eventually experiment with using it as a kind of tutorial.