Patreon Rejects Fair Use for AI Training, Demands Creator Pay
Austin, Texas – At the South by Southwest (SXSW) conference this week, Patreon CEO Jack Conte publicly dismissed the “fair use” argument employed by artificial intelligence companies regarding the use of copyrighted material for training their models, asserting It’s a “bogus” claim. Conte’s criticism centers on the disparity between how AI firms treat large rights holders and individual creators.
Conte, who founded Patreon in 2013 after struggling to monetize his own music on YouTube, has built a platform supporting over 3 million monthly active users. His remarks at SXSW positioned the current debate over AI and copyright as a recurring pattern of disruption faced by creators in the digital age, comparing it to shifts in the music industry and the rise of short-form video. Despite acknowledging the inevitability of AI, Conte insists creators must be compensated for the use of their function.
“I run a frickin’ tech company,” Conte stated, emphasizing that his position isn’t anti-AI but rather a defense of creator rights. He argued that the legal justification for utilizing creators’ content for AI training falls apart when considering the substantial licensing deals AI companies negotiate with established entities like Disney, Condé Nast, Vox, and Warner Music. “If it’s legal to just use it, why pay?” he asked rhetorically.
The core of Conte’s argument, echoed in statements reported by TechCrunch and AOL, is that AI companies are applying a double standard. They claim “fair use” to justify leveraging the work of individual creators while simultaneously recognizing the value of copyrighted material held by large corporations through paid licensing agreements. This inconsistency, Conte suggests, undermines the legitimacy of the “fair use” defense.
The legal doctrine of “fair use” allows limited use of copyrighted material without permission for purposes such as criticism, commentary, news reporting, teaching, scholarship, or research. AI companies have frequently cited this doctrine to justify using vast datasets of copyrighted content to train their models, arguing that the resulting AI systems represent a “transformative” use of the original material. Yet, Conte contends this argument is disingenuous given the parallel practice of securing paid licenses from major rights holders.
Conte’s statements come as the debate over AI and copyright intensifies. The issue has drawn scrutiny from creators across various industries, who fear that their work is being exploited without adequate compensation or attribution. As of March 22, 2026, no formal response has been issued by major AI companies to Conte’s public criticism, and the legal status of AI training data remains a contested area.