In recent weeks, Adobe has faced a sharply negative reaction from users to the new version of its terms of service. In response, the company is now trying to correct the situation. Recently, Adobe published a revised version of its terms of service agreement, clearly stating that it would not train artificial intelligence on user content stored locally or in the cloud.
The updated document now includes several categories defining Adobe’s access to user-generated content, with a specific section dedicated to generative AI. The terms of service explicitly state that the company’s software “will not use your on-premises or cloud content to train generative AI.” There is one exception: if the user’s material is submitted to the Adobe Stock site, the company reserves the right to use it to train its Firefly neural network.
Adobe’s director of strategic issues, Scott Belsky, emphasized that the new version of the document does not change the company’s existing policies. The previous lack of clarity in stating their position on AI training led to misunderstandings. “We have explicitly stated that we will not train generative AI on your content. This has always been our company policy. We have always been very clear about this, but we have never said it explicitly,” Mr. Belsky told The Verge.
Addressing User Concerns on Content Scanning
The revised document also addresses user concerns about the company’s scanning of content created under a non-disclosure agreement (NDA), adds NIXSolutions. Adobe assured users that it “does not scan or review” content stored locally on the user’s computer. However, it does automatically scan content in the cloud for illegal material. If the system detects such material or if the user participates in pre-release software testing or product improvement programs, the content in the cloud can be reviewed by a company employee.
We’ll keep you updated on any further developments regarding Adobe’s terms of service and user policies.