Challenges in Addressing IP Rights against Generative AI: Evidence Gaps and Redress Mechanisms in U.S. Cases
| Authors | |
|---|---|
| Year of publication | 2024 |
| MU Faculty or unit | |
| Citation | |
| Attached files | |
| Description | This research poster explores how intellectual property law responds to the rapid development of generative artificial intelligence in the U.S. legal system, focusing on litigation involving the use of copyrighted works in AI training. The study includes qualitative and quantitative analyses to understand the most frequently litigated claims and legal argumentation. Firstly, the study examines 21 cases involving generative AI, identifying direct infringement as the most common claim, followed by vicarious infringement and DMCA Section 1202 claims. Most cases originate from the Northern District of California and the Southern District of New York. Secondly, it analyzes five key procedural decisions, focusing on reasoning and evidence gaps. Direct copyright infringement claims are more resilient and often survive judicial scrutiny, while vicarious infringement and unjust enrichment claims frequently face dismissal due to lack of direct infringement or preemption by copyright law. Claims related to training new AI models have a better chance of surviving initial scrutiny. However, challenges arise in the post-training phases, including proving the infringement related to generated output and the existence of the AI model itself |
| Related projects: |