- Though generative AI may be new to the market, existing laws have significant implications for its use. Now courts are sorting out how the laws on the books should be applied.
- In many cases it poses legal questions that are still being resolved. For example, does copyright, patent, and trademark infringement apply to AI creations?
- Lawyers — who perhaps stand most to benefit from the muddied waters — offer advice tailored for creators, AI developers, and business users on what to do.
READ MORE: AI Art and Copyright Part 1: Do Androids Dream of Copyright Registration? (Dunlap Bennett & Ludwig)
READ MORE: AI Art & Copyright Part 2: Artificial Intelligence or Artfully Infringing? (Dunlap Bennett & Ludwig)
The meteoric rise of AI applications has left the industry wondering how this technology will interact with copyright law… and whether the law can keep up. This is largely uncharted territory, but here’s legal advice for the developers of AI tools and artists working with them.
There are two primary questions to consider about AI art. The first is, “Can AI art be copyrighted?” The other question surrounds the legal status of artists who claim to have had their art stolen (generously called “sampled”) to supply the data for AI diffusion models.
Thuan Tran, associate at Dunlap Bennett & Ludwig, answers the first question, stating that the US Copyright Office will reject a request to allow an AI to copyright a work of art. This is because it will not register works produced by a machine “or mere mechanical intervention” from a human author.
Courts interpreting the Copyright Act, including the Supreme Court, have consistently restricted copyright protection to “the fruits of intellectual labor” that “are founded in the creative powers of the [human] mind.”
However, this interpretation is being tested. In a case currently before the Supreme Court, artist Kris Kashtanova is contesting a decision by the Copyright Office not to register a copyright for graphic novel that she created using an AI.
Kashtanova is emphasizing in how she “engaged in a creative, iterative process” that involved multiple rounds of composition, selection, arrangement, cropping, and editing for each image in her work, which makes her the author of the work.
“While the outcome of the proceeding is not yet finalized and Kashtanova has a chance to appeal its decision, many are eagerly awaiting what may be very precedential for the future of AI art.”
The second question is also taken up by Tran, and is also being framed in the court of law. There are several cases of artists suing generative AI platforms for unauthorized use of their work
Image licensing service Getty, for example, has filed a suit against the creators of Stable Diffusion alleging the improper use of its photos, both violating copyright and trademark rights it has in its watermarked photograph collection.
The outcome of these cases is expected to hinge on the interpretation of the fair use doctrine. This is the legal concept that allows for the use of copyrighted material without permission from the copyright holder, in certain circumstances.
Tran explains that Fair use is determined on a case-by-case basis, and courts consider four factors: (1) the purpose and character of the use; (2) the nature of the copyrighted work; (3) the amount and substantiality of the portion used; and (4) the effect of the use upon the potential market for or value of the copyrighted work.
“One argument in favor of AI-generated art falling under fair use is that the use of copyrighted material by AI algorithms is transformative,” he says. “Transformative use is a key factor in determining fair use. It refers to the creation of something new and original that is not merely a copy or imitation of the original work.”
AI algorithms create new works by processing and synthesizing existing works, resulting in a product that could be considered distinct from the original. “As a result, AI-generated art can be seen as a form of transformative use, which would weigh in favor of fair use,” Tran says. “On the other hand, this argument is not without its limitations. Many argue that AI-generated art is simply a recombination or manipulation of existing works, without adding significant creative output. “
There is also the larger philosophical debate as to whether a machine can give “creative input” to its work. In such cases, it may be more difficult to argue that the use of copyrighted material is transformative and subsequently falls under fair use.
All this uncertainty presents a slew of challenges for companies that use generative AI. There are risks regarding infringement — direct or unintentional — in contracts that are silent on generative AI usage by their vendors and customers.
The Harvard Business Review gives some advice for AI vendors, their customers, and artists.
“AI developers should ensure that they are in compliance with the law in regards to their acquisition of data being used to train their models,” advise Gil Appel, Assistant Professor of Marketing at the GW School of Business, Juliana Neelbauer, partner at law firm Fox Rothschild LLP, and David A. Schweidel, Professor of Marketing at Emory University’s Goizueta Business School. “This should involve licensing and compensating those individuals who own the IP that developers seek to add to their training data, whether by licensing it or sharing in revenue generated by the AI tool.”
Developers should also work on ways to maintain the provenance of AI-generated content, which would increase transparency about the works included in the training data. This would include recording the platform that was used to develop the content, tracking of seed-data’s metadata, and tags to facilitate AI reporting, including the specific prompt that was used to create the content.
“Developing these audit trails would assure companies are prepared when customers start including demands for them in contracts as a form of insurance that the vendor’s works aren’t willfully, or unintentionally, derivative without authorization.
“Looking further into the future, insurance companies may require these reports in order to extend traditional insurance coverages to business users whose assets include AI-generated works.
When it comes individual content creators and brands, the onus is on them to take steps to protect their IP.
Stable Diffusion developer Stability.AI, for example, announced that artists will be able to opt out of the next generation of the image generator.
“But this puts the onus on content creators to actively protect their IP, rather than requiring the AI developers to secure the IP to the work prior to using it — and even when artists opt out, that decision will only be reflected in the next iteration of the platform. Instead, companies should require the creator’s opt-in rather opt-out.”
According to Appel, Neelbauer and Schweidel, this involves “proactively looking for their work in compiled datasets or large-scale data lakes, including visual elements such as logos and artwork and textual elements, such as image tags.”
Obviously, this could not be done manually through terabytes or petabytes of content data, but they think existing search tools “should allow the cost-effective automation of this task.”
Content creators are also advised to monitor digital and social channels for the appearance of works that may be derived from their own.
Longer term, content creators that have a sufficient library of their own IP on which to draw “may consider building their own datasets to train and mature AI platforms.”
The resulting generative AI models need not be trained from scratch but can build upon open-source generative AI that has used lawfully sourced content. This would enable content creators to produce content in the same style as their own work with an audit trail to their own data lake, or to license the use of such tools to interested parties with cleared title in both the AI’s training data and its outputs.
Customers of AI tools should ask providers whether their models were trained with any protected content, review the terms of service and privacy policies, “and avoid generative AI tools that cannot confirm that their training data is properly licensed from content creators or subject to open-source licenses with which the AI companies comply.”
If a business user is aware that training data might include unlicensed works or that an AI can generate unauthorized derivative works not covered by fair use, a business could be on the hook for willful infringement, which can include damages up to $150,000 for each instance of knowing use.
Consequently, businesses should evaluate their transaction terms to write protections into contracts. As a starting point, they should demand terms of service from generative AI platforms that confirm proper licensure of the training data that feed their AI.
Appel, Neelbauer and Schweidel add that they understand the real threat of generative AI to part of the livelihood of members of the creative class, “at the same time both creatives and corporate interests have a dramatic opportunity to build portfolios of their works and branded materials, meta-tag them, and train their own generative-AI platforms that can produce authorized, proprietary, (paid-up or royalty-bearing) goods as sources of instant revenue streams.”