IP Rights and AI Regulation: Finding the Right Balance

I’m excited to share a new article I recently published examining the relationship between intellectual property rights and artificial intelligence regulation, recently published in Volume 89, Issue 3 of the Missouri Law Review. Before diving into the substance, I want to express my sincere gratitude to the Missouri Law Review editors and staff, not only for their exceptional editorial work on this article but also for their tremendous efforts in organizing our March 2024 symposium that I co-organized on “AI and Society: Government, Policy, and the Law.” The symposium brought together dozens of leading scholars to explore the pressing challenges of AI governance, and those discussions helped shape many of the ideas presented in the piece.

• Dennis Crouch, Using Intellectual Property to Regulate Artificial Intelligence, 89 MO. L. REV. 781 (2024)

The article builds on foundational work of scholars, including Amanda Levendowski, who has explored how copyright law can address AI bias, as well as Mark Lemley and Bryan Casey, whose research on “fair learning” has shaped our understanding of AI training and copyright.  My article argues that we should be cautious about viewing IP as a primary regulatory tool.

The article advances two main claims. First, although IP rights help guide innovative behaviors in AI development, they aren’t well-suited as direct regulatory mechanisms. As Rebecca Wexler compellingly argued in her work on AI in criminal justice systems, traditional IP frameworks often lack the necessary levers to effectively regulate issues like algorithmic bias, privacy violations, or the concentration of power in the tech industry. Nicholson Price‘s work on black-box medicine and algorithmic accountability further reinforces this point, highlighting how IP protections can actually impede necessary oversight and transparency.

The tension between trade secrecy and transparency is particularly acute in AI regulatory discussions. While companies have legal interests in protecting their proprietary information, excessive trade secret claims can prevent meaningful oversight of AI systems that significantly impact people’s lives. As Christopher Morten has written, this creates a problem where the very information needed for effective oversight is shielded from public view by IP protections.  The article also examines recent regulatory developments, including the EU’s AI Act and various U.S. proposals, considering how these frameworks interact with existing IP regimes – expanding one the work done by my colleague Renee Henson.

I’m particularly indebted to the Missouri Law Review editorial team for their rigorous feedback and attention to detail throughout the publication process.

I mentioned Prof. Henson above. Her article from the same symposium examines whether the EU’s comprehensive AI Act could serve as a model for regulating emerging technologies in America, analyzing both the Act’s provisions and current U.S. regulatory proposals. Renee Henson, Bridging the Divide: Does the EU’s AI Act Offer Code for Regulating Emergent Technologies in America?, 89 MO. L. REV. 847 (2024).

The whole issue can be seen here: https://scholarship.law.missouri.edu/mlr/

I want to mention two other articles from the symposium that I really enjoy:

• Prof. Michael Siebecker published a corporate governance article arguing that existing corporate governance principles are inadequate to handle AI’s growing role in corporate decision-making. He calls for reconceiving basic corporate governance structures, particularly regarding corporate personhood and fiduciary duties. Michael R. Siebecker, Reconceiving Corporate Rights and Regulation in the AI Era, 89 MO. L. REV. 941 (2024).
• A really great collaborative team published their article focusing on legal issues arising from the complexity of the various data, security, and AI models within EU law. Iain Nash, DeBrae Kennedy-Mayo, Peter Swire & Annie Antón, Legal Issues in Reconciling Data Protection, AI, and Cybersecurity under EU Law, 89 MO. L. REV. 871 (2024).

Finally, I want to mention two of my IP-focused law students whose student notes are also published in this same Missouri Law Review article, Tom Langdon and Tim Knight.

• Thomas Langdon’s carefully drafted article on AI and antibody patents argues that artificial intelligence could help pharmaceutical companies satisfy the written description and enablement requirements for antibody genus claims in two ways: by improving the quality of written descriptions through generating more working examples, and by increasing the level of ordinary skill in the art. Thomas R. Langdon, Artificial Intelligence and Antibody Genus Claims, 89 MO. L. REV. 1031 (2024).
• Tim Knight’s article on copyright law examines the circuit split over claim accrual and recovery periods, arguing that courts should adopt the Discovery Accrual Rule while abandoning the three-year lookback period for recovery to better align with the spirit of copyright protection in the digital age. Tim Knight, Resolving the Incompatibility of Claim Accrual and Recovery in Copyright Law, 89 MO. L. REV. 1009 (2024). The article’s analysis proved prescient, as the Supreme Court largely adopted this approach in its May 2024 Warner Chappell Music v. Nealy decision, though it left open the broader question of whether the Discovery Accrual Rule should apply to copyright cases at all.

Original source can be found here.