Adobe users just now getting upset over content scanning allowance in Terms of Use
Update that triggered web rage didn't really change much – Photoshop slinger has always been able to scan your stuff
Updated A practically ancient terms-of-use update has landed Adobe in freshly boiled water over how the Photoshop giant gives itself the right to review user content stored in its cloud.
The issue kicked off online yesterday when video game concept artist Sam Santala pointed out the February update in a post on X. Santala asked if he was reading the paragraph right, seemingly dumbfounded by the implication.
"I can't use Photoshop unless I'm okay with you having full access to anything I create with it, INCLUDING NDA work," Santala posted, tagging Adobe for an explanation. A corporate representative did respond, but first – the offending passages.
All your art is belong to us?
A look at Adobe's General Terms of Use shows that they were last updated February 17, 2024, with changes made to two sections: 2.2 and 4.1, paragraphs which explain how Adobe will access user content and how it defines said content, respectively. Those are the paragraphs that have been called out online by Santala and other folks unhappy with the update.
The updated sections state that Adobe may access content through automated and manual methods "but only in limited ways, and only as permitted by law." That could include automated scanning of user content "using techniques such as machine learning in order to improve our services and software and the user experience," Adobe noted.
That passage might be the protesters' sticking point; some of them have threatened to switch to a competitor or open source product. Folks are upset at what looks like Adobe trying to give itself overly broad rights to use people's work however it sees fit, though the phrasing is that Adobe gives itself these boilerplate rights so that its products can work with people's materials.
As to what content Adobe will scan? Well, if it's uploaded to Adobe systems, it appears to be fair game.
"'Content' means any text, information, communication, or material, such as audio files, video files, electronic documents, or images, that you upload, import into, embed for use by, or create using [Adobe] services and software," the InDesign maker states in section 4.1. And yes, that includes Document Cloud services and software, too.
In a page linked from the terms covering Adobe's content analysis in greater detail, the software house said it mainly analyzes content to improve services, though Adobe's terms do note it "may use available technologies, vendors, or processes, including manual review, to screen for certain types of illegal content," including CSAM and "patterns of activity that indicate spam or phishing."
Content stored on a local PC with Adobe software installed won't be scanned, the biz said.
- Adobe sells fake AI-generated Israel-Hamas war images – then the news ran them as real
- Adobe will use your work to train its AI algorithms unless you opt out
- Adobe has 'no plans' to invest in XD despite failed Figma buy
- Canva acquires Affinity, further wounding a regulator-bruised Adobe
So yes, Adobe can scan your stuff, but those upset about changes to the terms back in February - changes that were largely minor, made to largely add clarifying language for the first time since 2022 - should realize the language has barely changed. In other words, Adobe's been able to use and scan your stuff the exact same way for some time, for many years in fact, by the looks of it.
That, and there's an opt-out, albeit a limited one.
Customers on personal Creative and Document Cloud accounts can disable content analysis - at least for the purposes of product improvement and development - using steps explained on the content analysis FAQ page linked above. Business-level accounts have such features controlled by the terms of their contract with Adobe.
Scott Belsky, Adobe's chief strategy officer and EVP of design and emerging products, responded to Santala's concerns on X, repeating his mantra that Adobe doesn't train any generative AI models on customer content, unless customers offer it up for cash, that is.
But Belsky's words weren't all reassurance, and make it clear that opt-out is limited.
"There are probably circumstances (like indexing to help you search your documents, updating components used from CC libraries across your documents, among others) where the company's terms of service allow for some degree of access," Belsky said.
Adobe, which warned late last year that it could face large fines over its subscription cancellation practices, didn't respond to questions for this story. ®
Updated to add on June 7
While Adobe didn't respond to questions from The Register, it did publish a blog post yesterday to address the matter, reiterating that it wouldn't use customer content to train AI.
"The focus of this update was to be clearer about the improvements to our moderation processes that we have in place," Adobe said. "Given the explosion of Generative AI and our commitment to responsible innovation, we have added more human moderation to our content submissions review processes."
Adobe said it retains a limited license to access user content "solely for the purpose of operating or improving the services and software and to enforce our terms and comply with law."