Hidden access: How much control does Adobe have over your data?
Adobe has already been at the center of public criticism on several occasions due to controversial business practices.
In June of this year, the company updated its General Terms and Conditions (GTC), giving itself the right to both automated and manual access to content stored by users in the Creative Cloud and Document Cloud. This caused considerable concern regarding data protection and the confidentiality of sensitive information.
So far, the change only affects the US edition. The original section 2 on data protection stated: “We will only access, view or listen to your content in limited ways.” In the revised version of the T&Cs, the text has now been changed: “We may access, view or listen to your content through both automated and manual methods, but only in limited ways.”
The new guidelines allow both automated checking by algorithms and manual analysis by Adobe employees. Machine learning analyses of the content are also planned.
The dark side of generative AI: a professional photographer speaks out
At Eyebase, we deal extensively with the management, distribution and rights management of digital assets, paying particular attention to data protection and data security. A recent case involving our customer, professional photographer Andreas Knurr, brought to light important questions about data protection in connection with Adobe's AI-supported functions.
The professional photographer wanted to try out the new edge enhancement feature in Adobe Camera Raw (ACR) module V17, which has only been available since this week. To activate this function, he first had to enable it as a technology preview in the ACR settings. After restarting Photoshop, a new option appeared when cropping: “Enable Expand”. To use the function, it is necessary to accept the specific conditions for AI once.
Images that are uploaded for editing are first checked to ensure that they comply with the previously accepted terms of use. In this case, the image shown below was rejected by the Adobe AI because it allegedly violated the terms of use. This prevented the professional photographer from continuing to edit the image.
These restrictions and the associated terms of use raise some data protection issues. The Generative AI Acceptable Use Policy, which Adobe last updated on May 10, 2024, states: “We may take action with respect to your Adobe account if we discover content or behavior that violates this policy.” This provision makes it clear that Adobe reserves the right to comprehensively monitor and, if necessary, restrict the use of its AI features if content is considered problematic.
Adobe requires that the Input and Output be used in accordance with the then-current version of the Terms and also reserves the right, under the current Terms, to throttle, restrict, disable, suspend or terminate your right to use the Generative AI features at any time in its sole discretion without prior notice.
A questionable outlook: Is Adobe on a dangerous course?
It is therefore primarily about detecting illegal or abusive material that goes against Adobe's usage guidelines. But even more worrying is the potential access not only by automated systems, but also by Adobe employees. The software manufacturer thereby secures potential access to unpublished projects, such as logos and image material for upcoming media productions.
The simplest solution would be not to store sensitive data online in the Creative Cloud or Document Cloud, but only on local data carriers to prevent Adobe from accessing it. However, some tools require an online connection. In particular, certain Photoshop functions that are based on artificial intelligence require an active internet connection - these include the “Neural Filter”, the Liquify filter, the AI-supported selection and also the generative AI tools.