If you use artificial intelligence to create something, do you own it?
While the courts work through this and other AI questions as the technology grows, early adopters of large language models and its AI stack likely will have to operate without clear regulatory guidance and end up in difficult business positions when regulators catch up to the technology. The intellectual property legal community, however, is making strides to keep up with technology breakthroughs.
Ownership of AI assisted IP creation
One of the most pressing questions is the ownership of intellectual property created with AI assistance, since current U.S. laws and courts only recognize human or legal entities as owners of various types of IP.
This recognition of ownership is rooted in the definition of inventorship in patent law and authorship in copyright law, where only a natural person is entitled to inventorship or authorship for patent and copyright protection purposes. A trade secret belongs to the entity taking necessary steps to protect it.
The ownership of AI-related IP that is solely created by humans, such as a patented AI system, copyrighted source code, or a trade secret data set still fall within the current legal system. However, AI may be able to create, or help create inventions that would be patentable or copyrightable if created by humans. In such cases, the AI system would itself have no right to the invention or the work of authorship under the current law.
The U.S. Copyright Office has canceled copyright protection it granted in 2022 to Kristina Kashtanova for her comic book Zarya of the Dawn – a book created by feeding text prompts to Midjourney, an artificial intelligence image generator. In March, the U.S. Copyright Office issued a regulation clarifying its practices for examining and registering works that contain material generated by the use of artificial intelligence technology. The guidance requires the copyright applicant who uses AI technology in creating a work to disclose the inclusion of AI-generated content in a work submitted for registration and provide a brief explanation of the human author’s contribution to the work.
The recent release of OpenAI’s rich developer tools will stimulate a wave of new technology startups that rely on ChatGTP. OpenAI has announced that it will protect its business customers against copyright infringement lawsuits, following the lead of Microsoft, Google and Amazon, who already offer customers such protection in relation to their generative AI products. This relief some pressure. But the question largely remains as how to assert the intellectual property rights of these new businesses and technology innovation. That includes obtaining ownership, licensing rights, and enforcement, which all will be evolving with the AI technology and its implications in the legal landscape.
What can an early AI adopters do?
There are ways, however, that a business can work to mitigate the risk while using AI assistance for intellectual property development. A company can obtain a freedom to operate (FTO) opinion before launching any AI-related product or service. A written FTO opinion may serve as evidence in later IP infringement litigation disproving willful or induced infringement – though it isn’t a legal guarantee.
A company can also include an indemnification provision in any license agreement for another party’s IP with AI element to reduce their patent or copyright infringement liability risk.
With the government and courts still catching up to the advances made in AI, it could be years before final regulations are established concerning innovation ownership issues. But as your company advances using AI, it’s always good to safeguard your creations before they’re illegal five years from now. If you have any questions on the effect of your innovations if you use AI, please contact a member of our Intellectual Property Group.
