Adobe said it would update the language in its terms of service to reflect that it doesn't rely on user content to train its generative AI tools. This comes after the company recently updated its terms of service. The terms of service state that Adobe “may access, display or view your content using both automated and manual methods,” but only in limited ways permitted by law. This includes purposes such as responding to feedback and support requests, detecting and preventing fraud and other legal and technical issues, and screening for illegal content.
Another part of the updated terms of service also says that in order for Adobe to operate and improve its services and software, users grant Adobe a license to “use, reproduce, publish, distribute, modify, create derivative works from, publicly perform, and translate the Content.”
Users were notified to re-accept the terms of service, sparking a huge backlash at X (formerly Twitter), as many argued that the access allowed the company to use customer data to train AI models, amounting to theft of users' work.
Background information:
The backlash against Adobe's updated terms of service highlights growing concerns and sensitivities about how companies use user data and content, especially in training artificial intelligence models. In a similar vein, Meta recently faced complaints from the Norwegian Consumer Council and privacy rights group None of Your Business (NOYB) over changes to its privacy policy that allowed public information shared by customers to be used to train AI models.
How did Adobe respond to the backlash?
The company published a blog post on June 6th about its updated terms of service, clarifying that it needs to access user content for the following purposes:
- Allows Adobe apps to perform the functions they are designed to perform, such as opening and editing files or creating previews of files for sharing.
- It offers cloud-based features like Photoshop Neural Filters, Liquid Mode, Background Removal, and more.
- Inspects content stored on Adobe servers for illegal content (such as child sexual abuse material) or other deceptive content or behavior (such as activity patterns indicative of spam or phishing).
The company released yet another clarification on June 10, emphasizing that users own their content and that Adobe does not use this content to train generative AI models. It also said it would add a similar statement to its terms of use to assure users that it has a legal obligation not to use their data for training purposes. It also said that the licenses required to operate its products use standard statutory copyrights, but that going forward it would explain with plain English examples what that means and why it's necessary.
Updated terms and conditions are expected to be published by June 18, 2024.
Read also:
Stay on top of tech news: Get a daily newsletter with MediaNama's top stories of the day delivered to your inbox before 9 a.m. Click here to sign up now.
