Unpacking the Complicated Relationship Between Adobe and Your Photos: How AI is Involved

Recently, a Krita developer spotted something unexpected in their Adobe Creative Cloud account settings: everyone had been opted into a “content analysis” program. This means that Adobe may use machine learning to analyze your content and improve its products and services. Some believed this meant the AI was ingesting images, but it’s more complicated than that.

During software installation, many offer an option to “share information with the developer,” which can send telemetry such as usage frequency and crash info. Unfortunately this isn’t always available (e.g., Microsoft’s Windows 10), sparking outrage among users who can’t turn it off.

It’s gross, but worse yet is introducing a new sharing method and automatically enrolling existing users in it. Adobe told PetaPixel that the content analysis has been around for 10 years – impressive if true, but likely not given no one seemed to be aware of it all this time. My guess is the policy has been in place some form or another and quietly changed over time.

The setting is clear: machine learning can be used to analyze your content, but not for training purposes. As the “learn more” link states:

For example, we may use machine learning-enabled features to help you organize and edit your images more quickly and accurately. With object recognition in Lightroom, we can auto-tag photos of your dog or cat. In Photoshop, machine learning can be used to automatically correct the perspective of an image for you.

Using machine learning, Lightroom and Photoshop can help you edit your images quickly, accurately and with ease. Object recognition in Lightroom auto-tags photos of your pets, while Photoshop’s ML capabilities can automatically adjust the perspective of a picture for you.

Adobe can use machine learning analysis to differentiate how many people are using Photoshop to edit images of people, landscapes, or other high-level metadata – information that could shape product decisions and priorities.

You may question if the images and analysis will be used to train AI models as part of product/service development.

Stock

Create a look like this. Credit: Adobe Stock

Adobe states clearly that no data stored in Creative Cloud accounts is used to train Generative AI features. While the legal precision of this statement may lead one to think otherwise, it does provide a clear explanation.

Examining the documentation shows that we aggregate your content with others’, then use it to improve our products and services via algorithm training.

Yes, it uses your content to train its algorithms – excluding its experimental Generative AI algorithms.

Adobe’s Photoshop Improvement Program (opt-in and documented here) provides a way to use your photos for training generative AI. In some cases, it also requires manual review.

If you value privacy, opt out of Adobe’s program and any others. Log in to the Privacy page to do so quickly and easily.

Avatar photo
Dylan Williams

Dylan Williams is a multimedia storyteller with a background in video production and graphic design. He has a knack for finding and sharing unique and visually striking stories from around the world.

Articles: 834

Leave a Reply

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