Adobe’s Policy Update: No Use of Artists’ Works to Train AI

MetaversePlanet June 20, 2024

0 1 minute read

Adobe has updated its terms of service policy, which was published in recent months and received backlash, and announced that it will not use users’ work to train artificial intelligence. However, the broken trust of the artists has not been restored.

Adobe’s new terms of service, updated last February, faced significant criticism. Adobe stated that users would be able to access its content “through both automated and manual methods” and that it would be able to access content through a program called “Firefly.” It also mentioned that it could use “techniques such as machine learning” to improve its tools.

Now, Adobe has released a new version of its terms of service agreement, clarifying that its AI features will not train on users’ content stored on devices or in the cloud. Additionally, it gave users the option to remove the permission to analyze their content.

Artists still don’t trust Adobe

Scott Belsky, Adobe’s chief strategy officer, explained that by machine learning, they mean non-productive AI tools such as Photoshop’s “Content Aware Fill.” While Adobe insists that the updated terms don’t give the company ownership of content and will never use users’ work to train Firefly, artists’ trust has already been broken.

The concern about the unauthorized use and commercialization of copyrighted works by generative artificial intelligence models is not new. Early last year, artist Karla Ortiz filed a lawsuit against Midjourney, DeviantArt, and Stability AI after her works were used under her own name in various generative AI models. The “fight” between artists and AI tools continues to grow on legal grounds.