Intellectual property (IP) protection is a tough area to navigate in the era of generative artificial intelligence (GenAI). For example, let’s say software developers apply GenAI to support with coding when writing a recent app, or you hire an artist to apply AI to generate graphics for a digital marketing campaign, but they spend a lot of time refining the vision and prompts. Do these types of work products benefit from IP protection?
The answer is: it depends.
IP is an theoretical product of the human intellect. Protecting intellectual property rights owners is significant because it enables creators and organizations to benefit from their investments in creativity and innovation. Copyrights, patents and other forms of intellectual property protection give your organization exclusive rights, which allows you to control revenue generation, distinguish you from the competition and prevent plagiarism of your work.
The problem is that with GenAI-generated content, it’s currently impossible to determine exactly how to do this because courts are lagging behind the technology. In this article, we will look at the current state of the law, how to protect GenAI’s work from an intellectual property perspective, and what works to avoid using GenAI in the first place.
Copyright and patents
Intellectual property laws were written long before GenAI existed. In the US, federal court rulings over the last few years (e.g. and ) seem to confirm that only human creators can benefit from copyrights or patents.
Currently, the Copyright Act requires human authorship and does not protect results obtained from recent forms of technology that operate without human intervention. The most favorable case for GenAI involved a comic book artist () who managed to retain copyright protection for his comic book narrative but for AI-generated images.
In the United States, it is even more obvious that the Patent Act requires inventors to be human. Of course, this does not prevent courts from determining whether sufficient human creativity or input actually gives rise to property, based on the specific facts of the future case. Does 50/50 human-GenAI software meet the criteria for copyright?
The answer is that we simply are not currently sure that we are at the threshold of human involvement. It’s wise not to rely too heavily on GenAI for things you absolutely need to copyright or patent. Using something like Copilot to support with routine coding tasks is probably fine, while the more artistic uses of GenAI are much more nebulous.
Different international perspectives
There is also a lack of international harmonization of intellectual property rules. Although countries generally reject patent applications when human inventors have not been identified, South Africa has allowed an artificial intelligence system to obtain patent protection due to the lack of a formal definition of “inventor” in its patent system.
Copyright protection has been less consistent across countries, both historically and recently. In a landmark case – unlike US courts – the Beijing Internet Tribunal ruled in November 2023 that a piece of AI-generated graphics uploaded to Xiaohongshu, one of China’s most popular social media platforms, could benefit from copyright protection. In this case, the plaintiff adjusted the prompts and parameters just enough to convince the court that the images reflected his human creativity and intellectual contributions.
According to the World Intellectual Property Organization (WIPO), South Africa, India, Up-to-date Zealand and the United Kingdom provide copyright protection for non-human computer-generated works, although protection may be narrower than for human-made works. Ukraine has introduced certain property rights to non-original objects generated by computer programs, such as non-original databases.
In other words: there appears to be a growing international appetite for protecting GenAI’s work product, but there is not yet a rubric that we can broadly apply.
However, there is one thing that can support protect your intellectual property better than copyrights and patents:
Trade secrets and contractual protection
A “trade secret” is confidential information that has economic value because it is not generally known or easily verified. It is also more compatible with GenAI results and may encourage organizations and their developers to more freely take advantage of GenAI’s benefits.
To obtain trade secret protection, organizations must demonstrate that they have tried to protect the relevant process or result. While this may not be useful for works of art displayed publicly, trade secret protection should arguably apply to AI-powered code, just as it has always applied to human-written code if it was privately developed and protected. Please note that the input used to train GenAI may still be subject to intellectual property infringement issues, but that would be the subject of another article.
Companies can commercially license GenAI-powered code as a trade secret while also incorporating greater protection into the licensing agreement itself, which could be in the form of a signed contract or even a click through to publicly available cloud solutions. The ability to apply license agreements to clearly define responsibilities and allocate risks for the company and the end user is extremely useful.
Takeaway for organizations
While courts and brilliant legal minds around the world are beginning to address intellectual property protection as it relates to GenAI, we are still quite early in the process.
Still, the current state of copyright and patent laws in the U.S. does not provide sufficient protection for AI-generated products. Organizations should carefully consider how to engage with GenAI and communicate clear principles to their employees and contractors.
There are three more shrewd things you can do to protect your GenAI work product. First, you should only consider using GenAI if intellectual property rights are not significant. This will significantly reduce the risk to your organization and the courts will begin to resolve these issues in a more concrete way. Second, the human role in the inventive or artistic process must be documented. Third, take advantage of trade secret protection and licensing agreements wherever you can. This could support recognize GenAI results as protected intellectual property in a way that copyright and patents simply cannot currently do.
about the author