Adobe Updates Terms of Service to Pledge No AI Training on User Data

Adobe has declared a firm commitment to not use user content to train its models.

In an update to its terms of use, the company said it has never trained its generative AI tools and models on user content and firmly pledges not to do so.

Adobe said it added the pledge to its terms of use to “reassure people that is a legal obligation.”

“Your content is yours and will never be used to train any generative AI tool,” the company said. “We will make it clear in the license grant section that any license granted to Adobe to operate its services will not supersede your ownership rights.”

Adobe offers a suite of generative AI-powered creative tools, including its image generation model, Firefly. The company said Firefly was trained on licensed content with permission and public domain images.

Businesses including Google and OpenAI use user data to update and train their AI services to perform better. Meta recently came under fire over its data handling practices for AI training. The Facebook and Instagram parent is the subject of 11 EU complaints over its AI training plans. 

Adobe said it does, however, use user data and content characteristics to “improve product experiences” including machine learning-powered features such as masking and background removal but nothing for generative AI. Users retain the option to opt out of this.

Related:Nvidia, Adobe and IBM to Let Third Parties Vet their AI Models

Adobe’s latest promise follows a January 2023 pledge to not use user data to train its AI models.

“If we ever allow people to opt-in for generative AI specifically, we need to call it out and explain how we’re using it,” Scott Belsky, the company’s chief product officer, said at the time.