Is Adobe using your photos to train its AI?

Recently, a developer from Krita discovered that Adobe had automatically enrolled all of their Creative Cloud users in a “content analysis” program without their knowledge or consent. According to the settings, the program “may analyze your content using techniques such as machine learning (e.g. for pattern recognition) to develop and improve our products and services.” Some have raised concerns that this means Adobe is using user-generated content to train their AI models. However, it’s not that simple.

Many software programs have an option to “share information with the developer,” which allows the company to collect telemetry data on how the app is used. This is often presented as an option during the installation process, but not always. Adobe has stated that this content analysis program has been in place for over a decade, however, the wording of the setting is unclear and it is not known if the company has been collecting this type of data for that long.

The purpose of the analysis is to improve Adobe’s products and services by using machine learning to more quickly and accurately organize and edit images, as well as to gain insight into how the software is being used by different types of users. This information can then be used to make product decisions and prioritize future development.

It’s important to note that the language in the setting does leave open the possibility that Adobe may be using the analyzed content to train AI models, but it is not made clear in the company’s statement.

Upon further examination of Adobe’s documentation, it appears that the company does indeed use user-generated content to train its algorithms for product improvement and development purposes. The content is first aggregated with other data before it is used to train the algorithms.

However, it is not clear if this process is also being used to train Adobe’s experimental Generative AI algorithms or not. The company does have a separate opt-in program called the Adobe Photoshop Improvement Program, which is documented and its purpose is specifically to use user’s content to train its AI models, but it’s possible that through some other means, user’s content is being used to train AI models that are not specified. It’s also possible for your content to be manually reviewed by Adobe employees for other purposes, which is a separate concern.

Even if it is not the case that Adobe is using your content to train its AI models, it’s important to be aware of the privacy implications and consider opting out of any programs like this if you value your privacy.

Leave a Reply

Your email address will not be published. Required fields are marked *