Couldn’t attend Transform 2022? Check out all the top sessions in our on-demand library now! Look here.
When OpenAI announced expanded beta access to DALL-E in July, the company offered paying subscribers full use rights to reprint, sell, and trade the images they create with its powerful text-to-image generator.
A week later, creative professionals in various sectors were already buzzing with questions. Top of the list: who owns images published by DALL-E, or for that matter, other AI-powered text-to-image generators, like Google’s Imagen? The owner of the AI training the model? Or the human urging the AI with words like “red panda wearing a black leather jacket and riding a motorcycle, watercolor style?”
In a statement to VentureBeat, an OpenAI spokesperson said, “OpenAI primarily retains ownership of the original image so we can better enforce our content policies.”
However, several creative professionals told VentureBeat that they were concerned about the lack of clarity about image ownership from tools like DALL-E. Some who work for major agencies or brands said these issues may be too uncertain to justify using the tools for high-profile client work.
MetaBeat will bring together thought leaders to offer advice on how metaverse technology will change the way all industries communicate and do business October 4 in San Francisco, CA.
Bradford Newman, who leads the machine learning and AI practice of global law firm Baker McKenzie, in his Palo Alto office, said the answer to the question “Who owns DALL-E images?” is far from clear. And, he stressed, legal consequences are inevitable.
“If DALL-E is adopted the way I think [Open AI] envisions, there will be a lot of revenue generated from using the tool,” he said. “And if you have a lot of players in the market and things that are at stake, you have a great chance of a lawsuit.”
Major interests are litigated for case-specific answers
Mark Davies, partner at Orrick, agreed that there are many open legal questions when it comes to AI. “What actually happens is when there are big interests, you litigate it,” he said. “And then you get the answers in a case-specific way.”
In the context of text-to-image generators and the resulting creations, the question is mostly about what constitutes “fair use,” he explained. Under U.S. copyright law, fair use is a “legal doctrine that promotes freedom of expression by permitting the unlicensed use of copyrighted works under certain circumstances.”
In a technology context, the most recent and largest example of a case was Google LLC v. Oracle America, Inc. from 2021. In a 6-2 decision, the Supreme Court ruled that Google’s use of Oracle’s code amounted to fair use under United States copyright law. The court therefore did not consider the question of whether the copied material is copyrighted.
A big lesson from that case was that these disputes will be settled by the courts, Davies emphasized. “The idea that we get a magical solution from another place is just not how the justice system really works,” he said.
However, he added that for a problem like that surrounding DALL-E image ownership to move forward, it often takes two parties that have a lot at stake because lawsuits are so expensive. “So it takes a fundamental disagreement about something very important for these rules to develop,” he said. And it has happened in the past, he added, with Morse code, with railways, with smartphones and the Internet.
“I think when you go through technological change, it feels unique and special,” he said. “But the industrial revolution has happened. It is arranged.”
Conflicting Statements from Open AI on DALL-E?
Still, some experts say Open AI’s statements about using DALL-E — that the company owns the images, but users can commercialize them — are confusing and contradictory.
Jim Flynn, a senior partner and managing director at Epstein Becker and Green, said it struck him as “a little giving with one hand, taking away with the other.”
The point is that both sides have pretty good claims and arguments, he noted. “Ultimately, I think the people who own this AI process are pretty good claiming that they would have some proprietary rights,” he said. “This image was created by the simple input of some basic 3rd party commands.”
On the other hand, an argument could be made that it is similar to using a digital camera, he added — where images are created but the camera manufacturers don’t own the rights to user photos.
In addition, if those who own text-to-image generators have image output, it would be “viscerally unsatisfactory” to many who believe that if they buy or license a process like DALL-E, they should own what they’ve made – especially if they paid for the right to use it in the exact same way the AI company promoted them to use it.
“If I were representing any of the ad agencies, or the ad agencies’ clients, I wouldn’t advise them to use this software to create a campaign, because I think the AI provider would have some claim to the intellectual property,” he said. “I’d like to negotiate something more final.”
The Future of DALL-E Image Ownership
While there are arguments on both sides of the DALL-E property issue, as well as many historical analogies, Flynn doesn’t necessarily think the law needs to be changed to address them.
“But will it change? Yes, I think so, because there are a lot of people, especially in the AI community, who have an interest that isn’t really related to copyright or intellectual property,” he said. “I think the interest in it is not driven by complex legal issues, but to push the issue of AI as a separate consciousness. Because so many other things in our society find their way to the courts to determine, that’s why these cases are here.”
Flynn predicts a shakeout, a new consensus on who owns AI-generated creations, which will be powered by economic forces that follow the law. “That’s what happened with things like email correspondence and legal privilege, and frankly, that’s what happened with the digital camera.”
He said he would tell customers that if they want to use AI-generated creations, it’s best to use a vendor like Shutterstock that offers a certain number of licenses for an annual fee.
“But the reality is that you also get big advertising agencies that are probably going to start developing their own advertising agencies.” [text-to-image AI]or license AI at the institutional level from an API provider to create ads,” he said. “And the ad agency will pay the creator of AI a certain amount of money and use that for clients. There are certainly models that fit this.”
The mission of VentureBeat is a digital city square for tech decision makers to gain knowledge about transformative business technology and transactions. Learn more about membership.