Jordana Sanft
416-596-1083
Paul-Erik Veel
416-865-2842
May 6, 2024

On the Horizon: Legal Complexities Intersecting Generative AI, Class Actions, and IP Law

The multifaceted nature of generative AI is bound to create legal complexities at the intersection of intellectual property law and class actions, as this emerging technology disrupts not only the tech landscape but the legal one too. 

Fostering innovation is a core tenet of intellectual property law in Canada. However, policy interests seek to balance the furtherance of technology and creativity with protecting the public. Class actions offer a process for advancing public interest by allowing representative plaintiffs to advance claims on behalf of an entire class of people. Class actions advancing consumer rights are commonplace. As generative AI becomes more mainstream, people will look to class action proceedings to address their grievances. As this happens, patent, copyright, trademark or other IP related allegations are also likely to become more present in such actions.

Consider, for example, an AI generated consumer product that fails to perform as intended or causes harm due at least in part to a patented technology. Affected individuals may seek recourse through a product liability class action. Similarly, disputes over data scraping or use of copyrighted materials to train AI or generate works are on the horizon. Creative works such as art and literature are already being used to train AI systems to create new works. Indeed, issues like these are already starting to appear in legal battles in the United States (see for example: PM v OpenAI LP; JL v Alphabet Inc; Andersen v Stability AI Ltd; and Getty Images (US), Inc v Stability AI, Inc. 

Discussion

Liability

A significant issue that class actions pertaining to generative AI will raise is establishing liability and accountability for the harm caused by AI. This involves grappling with questions about whether liability should and could fall on an AI system, its developers, users, or owners, and how to allocate responsibility between them. Liability becomes particularly challenging when multiple parties are involved in the development and deployment of the generative AI technology.

Proving harm and causation brings unique challenges in the context of generative AI. Demonstrating how the content created by generative AI harmed class members will be among the hurdles that need to be overcome if liability is to be made out. Additionally, establishing the causal link between an AI system and the alleged harm will pose its own challenges, likely requiring a host of experts with legal and technology appropriate backgrounds to advance strategies and present a cohesive case.

Disputes over IP ownership and inventorship could also arise in the context of a generative AI class action. The patent holder or the class could dispute liability on the basis of whether the generative AI is liable and to what extent, if any, its owners, creators, and authors/inventors may be held accountable. These legal battles could also assert complex issues of IP rights and enforcement like patent validity and infringement into the class action arena, with the further potential of impacting remedies like quantum of damages. Addressing such issues would require a deep understanding not only of generative AI and class action strategy but also of IP law, pushing the legal arena into unchartered territory.

Standing

Standing is another obstacle to overcome in class action cases directed at generative AI issues. For example, it remains unclear whether unauthorized use of the copyright works or personal data in AI models results in a legally recognizable harm, and if it is does, whether this constitutes an injury that is sufficient for a plaintiff to pursue their legal recourse and theory of the case. Would an exception to IP infringement such as fair use dispel an allegation of wrongdoing or harm? Depending on the circumstances, the use of copyrighted materials only for training purposes may not involve impermissible copying or substantial reproduction for commercial purposes as traditionally contemplated under copyright law. The viability of such claims remains to be seen.

Reverse Class Actions

Another interesting twist may be the rise of reverse class actions. In a conventional class action, a representative plaintiff sues a defendant on behalf of a class. In a reverse class action, an individual plaintiff sues a group of defendants who are alleged to have engaged in the same wrongful conduct. These reverse class actions require a representative defendant. There are several recent examples in the area of IP albeit, not involving AI (see Voltage Pictures v Salna and Seismotech IP Holdings Inc v John Does). These cases may unwittingly have laid the foundation for increase in class action proceedings before the Federal Court where generative AI and IP law are in issue. Although such actions could be brought in superior courts, the Federal Court with its strong IP capability and the ability to invalidate a patent in rem, creates an interesting choice of venue depending on the specifics of the issues in dispute.

The intersection of generative AI, class actions, and IP law presents a host of intricate issues that require careful consideration and expertise. To this end, there is a need for collaboration between experts and legal specialists working towards a holistic strategy that promotes innovation, advances IP rights, and grapples with class action standing and liability, among other issues, relating to the generative AI ecosystem.

Practical Tips

There remains a myriad of unresolved legal issues in this space, and it will be challenging for players operating in this area to take steps to completely avoid any risk of liability. Novel and disruptive technologies virtually always present some risk of liability by their nature. That is a risk that many players choose to accept in order to build novel technologies and products. 

That being said, while liability risk should not dissuade companies from working in the generative AI space, so too should companies be mindful of the liability risks posed and take steps to minimize them. For example:

Monitor the Landscape and Identify Best Practices

Participating in industry forums and knowledge sharing in the evolving landscape can provide important insight, help set the stage of new industry norms, and provide a means of mitigating risk. By proactively engaging in discussions assessing and addressing social, ethical, and legal implications, additional innovation fostering safeguards may be identified and adopted. Further, by identifying and tying to address possible risks, one can minimize the potential for liability and be better prepared to navigate class action and IP issues.

Engage Experts

When a potential issue arises, move quickly to get advice as to how to handle it. Engage with legal and technology experts early in the pre litigation and litigation context to obtain guidance on the complex legal and factual issues in your industry facing innovation and commercialization.

This is Part 3 of our 5-Part Series on AI in the Courtroom, which includes the below blogs.