Several artists have been sharing posts and links about a dispute brewing between artists and Adobe.
Photoshop has been a longtime tool for visual/digital artists. They pay a license to Adobe to use the company's software and services. But this month, a number of artists were sharing these posts, with their concern about unauthorized use of their name and artwork by Adobe.
The story goes back to earlier in 2023. Adobe made an update that didn't go unnoticed. Here's an excerpt from the January 2023 story "Is Adobe using your photos to train its AI? It’s complicated" by Devin Coldewey on Techcrunch. com.
"A
sharp-eyed developer at Krita noticed recently that, in the settings for their
Adobe Creative Cloud account, the company had opted them (and everyone else)
into a “content analysis” program whereby they “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 taken this to mean that it is
ingesting your images for its AI. And … they do. Kind of? But it’s not that
simple....
....Adobe told PetaPixel that this content analysis thing “is not new and has been in place for a decade.” If they were using machine learning for this purpose and said so a decade ago, that’s quite impressive, as is that apparently no one noticed that whole time. That seems unlikely. I suspect the policy has existed in some form but has quietly evolved...
......Adobe clarified that “Adobe does not use any data stored on customers’ Creative Cloud accounts to train its experimental Generative AI features.” That wording is clear enough, though it also has the kind of legal precision that makes you think they’re talking around something.
If you look closer at its documentation, it does indeed say: “When we analyze your content for product improvement and development purposes, we first aggregate your content with other content and then use the aggregated content to train our algorithms and thus improve our products and services.”
So it does use your content to train its algorithms. Perhaps just not its experimental Generative AI algorithms.
In fact, Adobe has a program specifically for doing that: the Adobe Photoshop Improvement Program, which is opt-in and documented here. But it’s entirely possible that your photos are, through one tube or another, being used as content to train a generative AI. There are also circumstances when it might be manually reviewed, which is a whole other thing.
Even
if it isn’t the case that Adobe is harvesting your creativity for its models,
you should opt out of this program and any others if you value privacy. You can
do so right here at the privacy page if you’re logged in."
This Techcrunch article was circulated again in August 2023, when several artists began posting examples of Adobe using their art without permission. This phase of the story is covered in “Artists complain of AI ‘copyright infringement’ on Adobe Stock” a 8.18.23 article by Daniel Piper on Creativebloq.com
Excerpt: “AI has proven an incredibly contentious topic in the world of art and design over the past couple of years, but one company had been particularly keen to emphasise its ethical approach to the tech. Adobe claims its Firefly AI model is only trained on Adobe Stock and commercially licensed imagery – but some artists are already accusing the brand of copyright infringement.
Several artists have observed that commercially available AI-generated imagery appears in Adobe Stock search results when their name is used as a prompt – even though they didn't create the art. And in some cases, the AI art appears to at least partially mimic the style of the artist. This has drawn the ire of several creatives on Twitter this week – and arguably isn't great optics for a brand so vocally committed to "doing the right thing" with AI.
The controversy comes two months after artist Kelly McKernan complained on Twitter that their name was clearly being used as a tag to sell AI-generated art on Adobe Stock (above). Since then, Wetterschneider has posted multiple examples of artists' names producing results often in the hundreds. These include Beeple, Victo Ngai, Michael Whelan and Loish.
"I'm against the use of my art in the databases
used for generative AI," Loish told Creative Bloq. "I think so-called
ethical AI stock images should not mimic the styles of artists who did not opt
in to the use of their art in such a way. It's copyright infringement, in my
eyes."
While it isn't clear whether the AI-generated artwork itself was trained on the artists' work, it seems the fact that their names are clearly linked to it – perhaps through tags or metadata – is enough for them to feel short-changed. As Wetterschneider puts it, "334 results for @Beeple - I don't know how he feels about charging $80 for art using his name as a selling point."
Adobe itself has replied to some of the comments on
Twitter. "Hi there, thanks for letting us know about this — it goes
against our generative AI content policy. We’ve escalated this with our team to
ensure it's addressed," the official Adobe account responded to Loish,
while Mike Chambers, senior director for community at Adobe, tweeted to
Wetterschneider, "Fyi, these should be removed from Stock now. These all
violate our Generative AI submission guidelines, but clearly were missed by our
moderators." Creative Bloq has reached out to Adobe for comment, and will
update this article on response.”
One case study is cover in this8.19.23 story on Gizmodo “Judge Rules Wholly AI-Created Art Can’t Get Copyright Protections” by Justin Carter. https://gizmodo.com/ai-made-art-not-copyrightable-us-copyright-office-1850754829
"For those in the entertainment industry (and creative fields more generally), AI has been an extremely divisive topic. As various websites like ArtStation have decided to basically let art created by the controversial technology persist on their platforms—either freely or with specific guidelines—Hollywood writers and actors are striking in part because they fear how studios could easily use the technology as a way to avoid paying them. Fortunately, intellectual property law is here to say that copyrights will continue to only apply to human-made works.
On Friday, US District Judge Beryl Howell reaffirmed that sentiment with her ruling, stating “human authorship is a bedrock requirement” for anything seeking a copyright. The declaration was made in regards to Stephen Thaler’s bid to have the government begin allowing copyright protections for AI-made works. Thaler, who serves as chief engineer of the neural network firm Imagination Engines, has been trying to push for protecting AI creations since 2018. At the time of his original application, the Copyright Office turned it down, saying “the nexus between the human mind and creative expression” was a key factor in a work being protected.
Howell, in this most recent ruling, reaffirmed that US copyright law “protects only works of human creation.” Further, she stated that creations made by “non-human actors” such as AI “need no incentivization with the promise of exclusive rights under United States law, and copyright was therefore not designed to reach them.” While she acknowledged that the copyright is meant to “adapt with the times,” she argued that anything seeking protection is required to have “an originator with the capacity for intellectual, creative, or artistic labor. Must that originator be a human being to claim copyright protection? The answer is yes.”
Hollywood studios have been gradually making steps towards bringing AI into their productions, and it’s likely they’ll continue to do so by having enough humans involved so the work technically qualifies as being authored by real people. The larger issue of AI’s use in creative fields hasn’t been resolved here—and given that artists are currently pursuing legal action about having their art being used to train AI, it’ll be awhile before there’s any kind of resolution in sight."
This
case is also featured in this link from Copyright Alliance.org
https://copyrightalliance.org/current-ai-copyright-cases-part-2/
Info graphic below by Michele Leivan.
No comments:
Post a Comment