Pirated AI Training Data and the Future of Recovery for Creatives

Authors

• Cassie Anderson

Abstract

The explosive growth of generative AI necessarily relies on near-limitless amounts of creative works, such as books, articles, music, and art for training. Data sourcing and training practices advanced by large technology companies have defaulted to piracy, which presupposes unfettered access to high-quality creative data at minimal cost. In a growing number of markets, generative AI presents an appealing alternative to human-made creative works—an alternative that directly competes with the very human creatives whose works are necessary to power such technology. Facing mass-scale piracy and creative markets diluted by AI content, creative professionals are in an even more precarious position than normal. Class action litigation against large tech companies involving training data piracy and copyright infringement is widespread but slow-going. This article examines potential avenues of recovery for aggrieved creatives, focusing mainly on class action litigation. A potential class member’s understanding of their rights is essential for class actions to work for creatives, rather than against them. Creative industries are in a volatile position. Thus, it is essential that professionals understand the scope and content of current litigation, as well as other preventative options moving forward, such as licensing designed specifically for training data.

First Page

114

Recommended Citation

Cassie Anderson, Pirated AI Training Data and the Future of Recovery for Creatives, 10 Bus. Entrepreneurship & Tax L. Rev. 114 (2026).
Available at: https://scholarship.law.missouri.edu/betr/vol10/iss1/7