Who Should Bear the Burden of AI-Generated Copyright Infringement?

Michael P. Goodyear’s recent article in “Issues in Science and Technology” explores the challenging legal terrain of copyright infringement in the age of generative AI. As AI systems like Claude generate creative works, the potential for unintentional copyright violations rises, prompting urgent legal questions. Goodyear addresses a pivotal concern: who should be held accountable when generative AI produces content that infringes on copyright?

Using a compelling hypothetical case, Goodyear introduces Shane, a college student who unwittingly creates a song with lyrics resembling Taylor Swift’s “Love Story” using an AI tool. Shane, unaware of the similarity, faces the risk of substantial legal consequences. This example illustrates the tension between creativity enabled by AI and the legal protections afforded to original works. Copyright laws, which grant creators exclusive rights over their intellectual property, are struggling to keep pace with the rapid advancements in AI technologies.

Goodyear highlights two key legal questions: Is training AI on copyrighted works without consent lawful? And who bears responsibility when AI generates infringing outputs? He critiques the current framework that focuses liability on either users or developers, arguing that neither approach effectively addresses the unique attributes of AI systems. Users often lack intent or knowledge of infringement, while developers, despite implementing safeguards, cannot fully predict AI behavior due to the “black box” nature of these systems.

Proposing a groundbreaking shift, Goodyear advocates treating AI systems as fictitious legal persons, directly liable for copyright infringements. This novel approach acknowledges AI’s semi-autonomous role in creating content and aligns with historical practices of conferring legal personhood on entities like corporations. By assigning primary liability to the AI system, courts could navigate the complexities of AI-generated works without unfairly penalizing users or stifling innovation among developers.

Secondary liability, under Goodyear’s framework, would address human culpability when users or developers act with intent or negligence. For example, developers who fail to mitigate known infringement risks or users who deliberately exploit AI for infringing purposes could still face consequences. This balanced approach seeks to deter bad actors while fostering a competitive and innovative AI industry.

The article also underscores the broader policy implications of redefining AI liability. Shifting focus from strict preemptive controls to ex-post measures like a “notice-and-revision” system would allow developers to address specific infringements after they occur. This model would reduce barriers for new market entrants while ensuring accountability for harmful outputs. Goodyear’s proposal challenges courts and lawmakers to rethink legal doctrines to accommodate AI’s growing autonomy and societal impact.

Goodyear’s analysis is a call to modernize copyright law in the face of transformative technology. By placing AI systems at the center of liability and refining secondary accountability for human actors, his framework offers a nuanced solution to a complex problem. This approach not only preserves the integrity of copyright protections but also supports the development of AI as a tool for societal advancement.

This blog post is a summary of Michael P. Goodyear’s article and does not guarantee accuracy or provide legal advice. Readers are encouraged to consult the original publication and seek professional guidance for specific legal concerns.