Adobe has clarified controversial shrinkwrap license terms, but the damage may have alread...
After a terms of service update that infuriated artists, and an initial statement that poured gasoline on the fire, Adobe has made a clear statement about its new use terms.

Adobe Creative Cloud logo
The last 48 hours have been tumultuous for Adobe. Early in the week of June 3, users of Adobe Creative Cloud pointed out that the new terms of service allowed Adobe to do whatever it wanted with users projects.
We saw that furor, and reached out to Adobe about it. Then, they issued an unclear statement on the matter, saying that the terms had always been this way.
"Adobe accesses user content for a number of reasons, including the ability to deliver some of our most innovative cloud-based features, such as Photoshop Neural Filters and Remove Background in Adobe Express, as well as to take action against prohibited content," the company said at the time. "Adobe does not access, view or listen to content that is stored locally on any user's device."
This didn't help, and didn't say anything about training AI at all. Especially considering that "Cloud" is the third word in the product name, and the statement didn't address that at all.
So, we wrote about it on Thursday after giving them a chance to explain themselves for three days, and getting nothing back. They finally said something concrete on Thursday night.
- Access is needed for Adobe applications and services to perform the functions they are designed and used for (such as opening and editing files for the user or creating thumbnails or a preview for sharing).
- Access is needed to deliver some of our most innovative cloud-based features such as Photoshop Neural Filters, Liquid Mode or Remove Background.
- For content processed or stored on Adobe servers, Adobe may use technologies and other processes, including escalation for manual (human) review, to screen for certain types of illegal content (such as child sexual abuse material), or other abusive content or behavior.
The Thursday post specifically says that "Adobe does not train Firefly Gen AI models on customer content" and "Adobe will never assume ownership of a customer's work."
This is all well and good. The latter wasn't really in question.
The former, however, is oddly specific. It's good that they made the statement that they aren't training Firefly with these materials.
What would have been better is a blanket statement saying that they won't use it for Firefly, and won't sell or license it to others to train their models. The other generative AI providers complain that properly licensing content to train models is too hard, so they shouldn't have to.
Adobe should be clear that they won't allow this.
The statement also doesn't address that the terms still seem to breach confidentiality agreements that artists may have signed, should they use cloud-based, well, anything, from Adobe. We'll see if that is addressed.
Based on our brief conversations this morning with attorneys that specialize in this kind of thing, it doesn't seem to be with Thursday's statement.
The company says that it will be clarifying the Terms of Use acceptance to reflect the details of Thursday's post. It's not clear when this is going to happen.
Adobe has an incredibly large platoon of lawyers, so it would have been better had they thought about this before.
The court of public opinion will get the final say
We're already seeing commentary on social media that Adobe got caught with their hands in the cookie jar, thus the statement. We're not sure about that, but regardless, the damage has been done.
For now, we can say that we've gotten a lot of emails overnight about our piece on Thursday. Most of them laid out multiple pain points with Adobe software above and beyond the terms of service update.
And, they claim to not be looking back after a switch. So, we'll see if this statement changes user sentiment.
Notably, Adobe didn't bother to email us about this statement. And, they have yet to respond to our emails about it on Monday, Tuesday, and Wednesday.
Read on AppleInsider


Comments
Apple made a nearly identical rights claim change in its Terms of Service at the end of March, worded in much the same way but far more vaguely. But no one noticed. Perhaps that's why Apple has never clarified what it means, either. My guess is the same as Adobe's, but like them, Apple ( and everyone else) needs to be clearer on it.
"Except to the extent prohibited by law, you hereby grant Apple a worldwide, royalty-free, perpetual, nonexclusive license to use the materials you submit within the Services and related marketing as well as to use the materials you submit for Apple internal purposes. Apple may monitor and decide to remove or edit any submitted material, including via automated content filters and/or human review."
There are too many questions surrounding how our data is being used across all LLM training and delivered AI services. Companies try to avoid discussing it unless the questions become too public. For Adobe they did.
One is known to not sell your data to third parties, and the other is not clear about it.
And yes, I agree that there are too many questions about LLM training. We'll see how Apple handles this.
By the way, what Adobe says doesn't matter at all; legal terms are legal terms, and providing an worldwide exclusive license have a precise meaning, that no amount of explication from Adobe can change. I don't think Adobe has bad intentions here, i think they just choose the bad lawyer to write the text. But the text is there, and it holds if you accept it.
"Content” means any information that may be generated or encountered through use of the Service, such as data files, device characteristics, written text, software, music, graphics, photographs, images, sounds, videos, messages and any other like materials.
You can make anything look bad when you take it out of context.
I'm not a lawyer, but I think it does, actually. It's pretty clear that if you put anything on their servers, they reserve the right to scan it for illegal activity, and if such activity is detected, escalate it for human review, and then on to law enforcement, etc., as necessary.
Nondisclosure agreements do not shield illegal activity. So if you're photoshopping kiddie porn and using Adobe's servers to do it, the breach of your NDA is the least of your problems.
If you're not operating a criminal conspiracy and your nondisclosure agreement says you will guarantee that no one will ever possibly be able to look at covered content, then you're going to do your work in your own little SCIF on a device that isn't connected to the internet.
I'm not an expert, but my bet is that most NDAs will say something more along the lines that you will take reasonable measures to assure the security of the covered content. In that case, posting drafts of your work on instagram to get your friends' opinions of it is a violation of your NDA. Storing it on a server that's scanned for illegal activity probably isn't, nor would it be a violation if someone breaks into your locked office and steals the material.
It's absolutely valid to ask questions about these things, but especially when the issue blows up publicly, it should be no surprise that it takes a few days to get an answer. Once it's gone off the rails, the engineers, executives and lawyers will all be in sweaty meetings thinking, rethinking and overthinking the response, because they know at that point whatever they say will be challenged with a "yeah, but what do you mean by..." response.
Also, didn’t know GatorGuy had taken a second “shill” job. Congrats!
https://www.adobe.com/privacy/policy.html#info-secure
We also disclose personal information to other companies in the Adobe family and with advertising and sales partners consistent with your choices.
https://www.apple.com/legal/internet-services/itunes/
Apple explicitly states it’s for internal purposes. I can imagine training scenarios, going over reviews, iBooks trends, etc.
The key word is internal.
Ai has become an excuse to plagiarize without penalty. In reality, it shouldn’t matter if it’s a mashup of stored content. That content was created by many others. They deserve to control or monetize their content. Not have it ripped off so a bunch of people lacking in those gifts can ask a chat bot to “make” it.