AI & RoboticsNews

Adobe Stock creators aren’t happy with Firefly, the company’s ‘commercially safe’ gen AI tool

Adobe’s stock soared after a strong earnings report last week — where executives touted the success of its “commercially safe” generative AI image generation platform Adobe Firefly. They say Firefly was trained on hundreds of millions of licensed images in the company’s royalty-free Adobe Stock offering, as well as on “openly licensed content and other public domain content without copyright restrictions.” On the Firefly website, Adobe says it is “committed to developing creative generative AI responsibly, with creators at the center.”

“We could not be more excited about our generative AI road map that will make Adobe products more accessible to an even larger universe of people, while dramatically enhancing productivity for existing customers,” said David Wadhwani, president, digital media business at Adobe.

But a vocal group of contributors to Adobe Stock, which includes 300 million images, illustrations and other content that trained the Firefly model, say they are not happy. According to some creators, several of whom VentureBeat spoke to on the record, Adobe trained Firefly on their stock images without express notification or consent.

While this is certainly an issue for other text-to-image generative tools such as  DALL-E 2, Stable Diffusion and Midjourney (which were trained on scrapes of imagery posted to the public web, including copyrighted imagery), it is particularly egregious for a company like Adobe, which has been deeply intertwined with the creative economy for decades, they say.

Now, Adobe Stock creators say Firefly’s popularity is making it far less likely that users will purchase stock images. According to Adobe, since its launch in March, Firefly beta users have generated over 200 million images using a variety of newly available tools and features such as text-to-image, generative fill and extend image. Photoshop users generated over 150 million images in just the first two weeks using the new generative fill feature powered by Firefly.

In addition, a flooding of gen AI images into Adobe Stock is cannibalizing the platform, the creators say. According to PetaPixel, Adobe Stock is currently the only major stock website accepting AI image submissions from contributors — including those generated in non-Firefly tools — and AI images are outperforming human-generated files on the site on many metrics (An Adobe spokesperson says “Adobe Stock respects the rights of third parties and requires all Stock contributors to comply with our terms, including those specific to the use of generative AI tools. You can find those terms here.”)

Dean Samed is a UK-based creator who works in Photoshop image editing and digital art. He told VentureBeat over Zoom that he has been using Adobe products since he was 14 years old, and has contributed over 2,000 images to Adobe Stock.

“They’re using our IP to create content that will compete with us in the marketplace,” he said. “Even though they may legally be able to do that, because we all signed the terms of service, I don’t think it is either ethical or fair.”

He said he didn’t receive any notice that Adobe was training an AI model. “I don’t recall receiving an email or notification that said things are changing, and that they would be updating the terms of service,” he said.

According to Eric Urquhart, a Connecticut-based artist who has a day job as a matte artist in a major animation studio, artists who joined Adobe Stock years ago could never have anticipated the rise of generative AI.

“Back then, no one was thinking about AI,” said Urquhart, who joined Adobe Stock in 2012 and has several thousand images on the platform. “You just keep uploading your images and you get your residuals every month and life goes on — then all of a sudden, you find out that they trained their AI on your images and on everybody’s images that they don’t own. And they’re calling it ‘ethical’ AI.”

Adobe Stock creators also say Adobe has not been transparent. “I’m probably not adding anything new because they will probably still try to train their AI off my new stuff,” said Rob Dobi, a Connecticut-based photographer. “But is there a point in removing my old stuff, because [the model] has already been trained? I don’t know. Will my stuff remain in an algorithm if I remove it? I don’t know. Adobe doesn’t answer any questions.”

The artists say that even if Adobe did not do anything illegal and this was indeed within their rights, the ethical thing to do would have been to pre-notify their Adobe Stock artists about the Firefly AI training, and offer them an opt-out option right from the beginning.

Adobe, in response to the artists’ claims, told VentureBeat by email that its goal is to build generative AI in a way that enables creators to monetize their talents, much as Adobe has done with platforms like Behance. It is important to note, a spokesperson says, that Firefly is still in beta.

“During this phase, we are actively engaging the community at large through direct conversations, online platforms like Discord and other channels, to ensure what we are building is informed and driven by the community,” the Adobe spokesperson said, adding that Adobe remains “committed” to compensating creators. As Firefly is in beta, “we will provide more specifics on creator compensation once these offerings are generally available.”

Hi there, thanks for your question. Adobe’s use of Stock content is covered by our Stock Contributor license agreement. We are developing a compensation model for Stock contributors and will share details once Firefly is out of beta. ^BT

Back in March, Adobe released Firefly at its annual conference, Adobe Summit. Similar to popular tools like DALL-E 2, Stable Diffusion and Midjourney, its biggest differentiators were its unique access to the massive number of images within Adobe Stock and a user interface that would allow people to use Firefly via Photoshop, Illustrator and other tools for commercial use.

Last week, Adobe also announced it will bring Firefly to enterprise users. It not only touted its “commercially-safe” approach, but said it also plans to provide enterprise customers with an indemnification against copyright claims for new imagery generated with Firefly, similar to what is currently in place for Adobe Stock.

While it stands by the safety of Firefly, “if a customer is sued for infringement, Adobe would take over legal defense and provide some monetary coverage for those claims,” a company spokesperson said.

Bradford Newman, who leads global law firm Baker McKenzie’s machine learning and AI practice in its Palo Alto office, said Adobe’s “commercially-safe” execution and indemnification offer is one of the first and the cleanest that he has seen — because Firefly was trained on Adobe Stock imagery provided by creators, and which Adobe says it has the ability to use for this purpose according to its Stock Contributor license agreement.

“It’s like having, in a way, a closed ecosystem,” he said. “What you’re warranty-ing is access to an ecosystem that’s trained and runs on a clean dataset, which as a solution has been discussed and contemplated for a while, but has never to my knowledge been fully executed at an enterprise level.”

Newman emphasized that he had not read Adobe’s agreement with Stock contributors and could not comment on it specifically. But Adobe’s Stock Contributor Agreement dated March 1, 2022 states: “You grant us a non-exclusive, worldwide, perpetual, fully-paid, and royalty-free license to use, reproduce, publicly display, publicly perform, distribute, index, translate, and modify the Work for the purposes of operating the Website; presenting, distributing, marketing, promoting, and licensing the Work to users; developing new features and services; archiving the Work; and protecting the Work.”

Legal experts say Adobe Stock artists and creators likely will not have the kind of legal leg to stand on that Adobe’s enterprise users will enjoy. Legal scholar Andres Guadamuz, a reader in intellectual property law at the University of Sussex in the U.K. who has been studying legal issues around generative AI, said that the language in Adobe’s Terms of Service tends to be very broad. “You give Adobe a license for perpetuity, for whatever medium shall be invented,” he said. “People don’t read those terms and conditions.”

In addition, he said that he doesn’t believe an image generated using a model is a derivative of the billions of images in the dataset — so it would likely not infringe on an artist’s copyright.

Newman agreed, adding that while he had not looked at the contracts Adobe Stock contributors signed, he did not think the artists’ argument was persuasive.

“As I understand it, they’re saying we’re fine with the stock images being used for someone to buy and iterate on in Photoshop, but if it’s used as a dataset for generative AI, somehow there’s an issue and we’re being ripped off,” he said.

But Nathaniel Bach, an attorney at Los Angeles-based Manatt, Phelps and Phillips who specializes in entertainment law, copyright and IP, pointed out that while he is not familiar with the Adobe Stock license, the current issues are part of an age-old conundrum around unanticipated technological use, such as Blu-Ray and DVDs and streaming. That is: Is future media covered by prior contracts?

“Courts have wrestled with this and come to different decisions depending on how widespread the language is and how much time has passed since the contract was entered into,” he told VentureBeat by phone. “So this sort of feels new again, with AI.”

Bach emphasized that while he doesn’t necessarily think Adobe’s actions are an overreach, he is sympathetic with the creators — he does a lot of artist advocacy work, particularly in the music space, he explained, where many agree that the industry needs to be careful about taking away the “lifeblood” of artists. “I think that one of the important things that’s happening now is that artists are speaking up and using their voices,” he said.

“We hear the artists’ concerns,” said the Adobe spokesperson, adding that as the company speaks with the community, “we are also hearing a great deal of excitement for what these new tools can mean in terms of their productivity, and the creativity it can unlock for creators of any skill level.”

But Dobi emphasized that this creativity can easily be simply a passable copy of another artist’s work — if an artist uses Firefly to create a standalone image through a prompt.

“I don’t know if you’ve looked at my stock photography, but I’ve spent the last 20 years photographing abandoned buildings across the Northeast and I’ve built up quite a library of images of it, I’ve had a book published, I just had a piece in the New York Times,” he explained. “Now I saw some AI artist [online] saying, ‘Show me your urban exploration photos built through AI, I built these through Adobe Firefly’ and I looked at these photos and they could pass as my photos, I wouldn’t question whether they were real photos unless you looked really closely. Someone using Firefly could easily put in a prompt with words like ‘mental asylum, symmetrical, natural light, peeling paint, textured walls, dirty floor,’ stuff like that.”

For example, the following is one of Dodi’s Adobe Stock photos:

And Dodi used prompts in Firefly to generate images that, while not identical, are certainly similar to his own work:

Samed said that Adobe Stock is “not a feasible platform for us to operate in anymore,” adding that the marketplace is “completely flooded and inundated with AI content.”

Adobe should “stop using the Adobe Stock contributors as their own personal IP, it is just not fair,” he said, “and then the derivative that was created from that data scrape is then used to compete against the contributors that [built and supported] that platform from the beginning.”

Dobi said he has noticed his stock photos have not been selling as well. “Someone can just type in a prompt now and recreate the images based off your hard work,” he said. “And Adobe, which is supposed to be, I mean, I guess they thought they were looking out for creators, apparently aren’t because they’re stabbing all their creators that helped create their stock library in the back.”

Urquhart said that as an artist in his mid-50s who also does analog fine art, he feels he can “ride this out,” but he wonders about the next generation of artists who have only worked with digital tools. “You have very talented Gen Z artists, they have the most to worry about,” he said. “Like if all of a sudden AI takes over and iPad digital art is no longer relevant because somebody just typed in a prompt and got five versions of the same thing, then I can always just pick up my paintbrush.”

From his perspective, Samed said, generative AI is “an arms race” using technology no one truly understands — and companies are moving too quickly and being reckless.

“The damage that’s going to be done is going to be unlike anything we’ve ever seen before,” he said. “I’m in the process of selling my company, I’ve got out — I don’t want to participate or compete in this marketplace anymore.”

Join top executives in San Francisco on July 11-12, to hear how leaders are integrating and optimizing AI investments for success. Learn More


Adobe’s stock soared after a strong earnings report last week — where executives touted the success of its “commercially safe” generative AI image generation platform Adobe Firefly. They say Firefly was trained on hundreds of millions of licensed images in the company’s royalty-free Adobe Stock offering, as well as on “openly licensed content and other public domain content without copyright restrictions.” On the Firefly website, Adobe says it is “committed to developing creative generative AI responsibly, with creators at the center.”

“We could not be more excited about our generative AI road map that will make Adobe products more accessible to an even larger universe of people, while dramatically enhancing productivity for existing customers,” said David Wadhwani, president, digital media business at Adobe.

But a vocal group of contributors to Adobe Stock, which includes 300 million images, illustrations and other content that trained the Firefly model, say they are not happy. According to some creators, several of whom VentureBeat spoke to on the record, Adobe trained Firefly on their stock images without express notification or consent.

While this is certainly an issue for other text-to-image generative tools such as  DALL-E 2, Stable Diffusion and Midjourney (which were trained on scrapes of imagery posted to the public web, including copyrighted imagery), it is particularly egregious for a company like Adobe, which has been deeply intertwined with the creative economy for decades, they say.

Event

Transform 2023

Join us in San Francisco on July 11-12, where top executives will share how they have integrated and optimized AI investments for success and avoided common pitfalls.


Register Now

Now, Adobe Stock creators say Firefly’s popularity is making it far less likely that users will purchase stock images. According to Adobe, since its launch in March, Firefly beta users have generated over 200 million images using a variety of newly available tools and features such as text-to-image, generative fill and extend image. Photoshop users generated over 150 million images in just the first two weeks using the new generative fill feature powered by Firefly.

In addition, a flooding of gen AI images into Adobe Stock is cannibalizing the platform, the creators say. According to PetaPixel, Adobe Stock is currently the only major stock website accepting AI image submissions from contributors — including those generated in non-Firefly tools — and AI images are outperforming human-generated files on the site on many metrics (An Adobe spokesperson says “Adobe Stock respects the rights of third parties and requires all Stock contributors to comply with our terms, including those specific to the use of generative AI tools. You can find those terms here.”)

Adobe Stock creators say it is unethical to train Firefly using their IP

Dean Samed is a UK-based creator who works in Photoshop image editing and digital art. He told VentureBeat over Zoom that he has been using Adobe products since he was 14 years old, and has contributed over 2,000 images to Adobe Stock.

“They’re using our IP to create content that will compete with us in the marketplace,” he said. “Even though they may legally be able to do that, because we all signed the terms of service, I don’t think it is either ethical or fair.”

He said he didn’t receive any notice that Adobe was training an AI model. “I don’t recall receiving an email or notification that said things are changing, and that they would be updating the terms of service,” he said.

According to Eric Urquhart, a Connecticut-based artist who has a day job as a matte artist in a major animation studio, artists who joined Adobe Stock years ago could never have anticipated the rise of generative AI.

“Back then, no one was thinking about AI,” said Urquhart, who joined Adobe Stock in 2012 and has several thousand images on the platform. “You just keep uploading your images and you get your residuals every month and life goes on — then all of a sudden, you find out that they trained their AI on your images and on everybody’s images that they don’t own. And they’re calling it ‘ethical’ AI.”

Adobe Stock creators also say Adobe has not been transparent. “I’m probably not adding anything new because they will probably still try to train their AI off my new stuff,” said Rob Dobi, a Connecticut-based photographer. “But is there a point in removing my old stuff, because [the model] has already been trained? I don’t know. Will my stuff remain in an algorithm if I remove it? I don’t know. Adobe doesn’t answer any questions.”

The artists say that even if Adobe did not do anything illegal and this was indeed within their rights, the ethical thing to do would have been to pre-notify their Adobe Stock artists about the Firefly AI training, and offer them an opt-out option right from the beginning.

Adobe, in response to the artists’ claims, told VentureBeat by email that its goal is to build generative AI in a way that enables creators to monetize their talents, much as Adobe has done with platforms like Behance. It is important to note, a spokesperson says, that Firefly is still in beta.

“During this phase, we are actively engaging the community at large through direct conversations, online platforms like Discord and other channels, to ensure what we are building is informed and driven by the community,” the Adobe spokesperson said, adding that Adobe remains “committed” to compensating creators. As Firefly is in beta, “we will provide more specifics on creator compensation once these offerings are generally available.”

Adobe released Firefly in March, focused on commercial use

Back in March, Adobe released Firefly at its annual conference, Adobe Summit. Similar to popular tools like DALL-E 2, Stable Diffusion and Midjourney, its biggest differentiators were its unique access to the massive number of images within Adobe Stock and a user interface that would allow people to use Firefly via Photoshop, Illustrator and other tools for commercial use.

Last week, Adobe also announced it will bring Firefly to enterprise users. It not only touted its “commercially-safe” approach, but said it also plans to provide enterprise customers with an indemnification against copyright claims for new imagery generated with Firefly, similar to what is currently in place for Adobe Stock.

While it stands by the safety of Firefly, “if a customer is sued for infringement, Adobe would take over legal defense and provide some monetary coverage for those claims,” a company spokesperson said.

Bradford Newman, who leads global law firm Baker McKenzie’s machine learning and AI practice in its Palo Alto office, said Adobe’s “commercially-safe” execution and indemnification offer is one of the first and the cleanest that he has seen — because Firefly was trained on Adobe Stock imagery provided by creators, and which Adobe says it has the ability to use for this purpose according to its Stock Contributor license agreement.

“It’s like having, in a way, a closed ecosystem,” he said. “What you’re warranty-ing is access to an ecosystem that’s trained and runs on a clean dataset, which as a solution has been discussed and contemplated for a while, but has never to my knowledge been fully executed at an enterprise level.”

Newman emphasized that he had not read Adobe’s agreement with Stock contributors and could not comment on it specifically. But Adobe’s Stock Contributor Agreement dated March 1, 2022 states: “You grant us a non-exclusive, worldwide, perpetual, fully-paid, and royalty-free license to use, reproduce, publicly display, publicly perform, distribute, index, translate, and modify the Work for the purposes of operating the Website; presenting, distributing, marketing, promoting, and licensing the Work to users; developing new features and services; archiving the Work; and protecting the Work.”

Legal experts say Adobe Stock artists and creators likely will not have the kind of legal leg to stand on that Adobe’s enterprise users will enjoy. Legal scholar Andres Guadamuz, a reader in intellectual property law at the University of Sussex in the U.K. who has been studying legal issues around generative AI, said that the language in Adobe’s Terms of Service tends to be very broad. “You give Adobe a license for perpetuity, for whatever medium shall be invented,” he said. “People don’t read those terms and conditions.”

In addition, he said that he doesn’t believe an image generated using a model is a derivative of the billions of images in the dataset — so it would likely not infringe on an artist’s copyright.

Newman agreed, adding that while he had not looked at the contracts Adobe Stock contributors signed, he did not think the artists’ argument was persuasive.

“As I understand it, they’re saying we’re fine with the stock images being used for someone to buy and iterate on in Photoshop, but if it’s used as a dataset for generative AI, somehow there’s an issue and we’re being ripped off,” he said.

But Nathaniel Bach, an attorney at Los Angeles-based Manatt, Phelps and Phillips who specializes in entertainment law, copyright and IP, pointed out that while he is not familiar with the Adobe Stock license, the current issues are part of an age-old conundrum around unanticipated technological use, such as Blu-Ray and DVDs and streaming. That is: Is future media covered by prior contracts?

“Courts have wrestled with this and come to different decisions depending on how widespread the language is and how much time has passed since the contract was entered into,” he told VentureBeat by phone. “So this sort of feels new again, with AI.”

Bach emphasized that while he doesn’t necessarily think Adobe’s actions are an overreach, he is sympathetic with the creators — he does a lot of artist advocacy work, particularly in the music space, he explained, where many agree that the industry needs to be careful about taking away the “lifeblood” of artists. “I think that one of the important things that’s happening now is that artists are speaking up and using their voices,” he said.

Creativity, or a passable copy? 

“We hear the artists’ concerns,” said the Adobe spokesperson, adding that as the company speaks with the community, “we are also hearing a great deal of excitement for what these new tools can mean in terms of their productivity, and the creativity it can unlock for creators of any skill level.”

But Dobi emphasized that this creativity can easily be simply a passable copy of another artist’s work — if an artist uses Firefly to create a standalone image through a prompt.

“I don’t know if you’ve looked at my stock photography, but I’ve spent the last 20 years photographing abandoned buildings across the Northeast and I’ve built up quite a library of images of it, I’ve had a book published, I just had a piece in the New York Times,” he explained. “Now I saw some AI artist [online] saying, ‘Show me your urban exploration photos built through AI, I built these through Adobe Firefly’ and I looked at these photos and they could pass as my photos, I wouldn’t question whether they were real photos unless you looked really closely. Someone using Firefly could easily put in a prompt with words like ‘mental asylum, symmetrical, natural light, peeling paint, textured walls, dirty floor,’ stuff like that.”

For example, the following is one of Dodi’s Adobe Stock photos:

And Dodi used prompts in Firefly to generate images that, while not identical, are certainly similar to his own work:

Adobe Stock “not a feasible platform for us to operate in anymore” 

Samed said that Adobe Stock is “not a feasible platform for us to operate in anymore,” adding that the marketplace is “completely flooded and inundated with AI content.”

Adobe should “stop using the Adobe Stock contributors as their own personal IP, it is just not fair,” he said, “and then the derivative that was created from that data scrape is then used to compete against the contributors that [built and supported] that platform from the beginning.”

Dobi said he has noticed his stock photos have not been selling as well. “Someone can just type in a prompt now and recreate the images based off your hard work,” he said. “And Adobe, which is supposed to be, I mean, I guess they thought they were looking out for creators, apparently aren’t because they’re stabbing all their creators that helped create their stock library in the back.”

Urquhart said that as an artist in his mid-50s who also does analog fine art, he feels he can “ride this out,” but he wonders about the next generation of artists who have only worked with digital tools. “You have very talented Gen Z artists, they have the most to worry about,” he said. “Like if all of a sudden AI takes over and iPad digital art is no longer relevant because somebody just typed in a prompt and got five versions of the same thing, then I can always just pick up my paintbrush.”

From his perspective, Samed said, generative AI is “an arms race” using technology no one truly understands — and companies are moving too quickly and being reckless.

“The damage that’s going to be done is going to be unlike anything we’ve ever seen before,” he said. “I’m in the process of selling my company, I’ve got out — I don’t want to participate or compete in this marketplace anymore.”

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.


Author: Sharon Goldman
Source: Venturebeat

Related posts
AI & RoboticsNews

H2O.ai improves AI agent accuracy with predictive models

AI & RoboticsNews

Microsoft’s AI agents: 4 insights that could reshape the enterprise landscape

AI & RoboticsNews

Nvidia accelerates Google quantum AI design with quantum physics simulation

DefenseNews

Marine Corps F-35C notches first overseas combat strike

Sign up for our Newsletter and
stay informed!