OpenAI’s Amici on “Market Dilution” and Copyright Fair Use

by Stephan Kinsella on November 6, 2026

Rose Esfandiari, “OpenAI Touts Briefs Backing Fair Use for AI Training,” IPWatchdog (Nov. 2, 2026).

Cross-motions for summary judgment on fair use are pending in the consolidated copyright suits against OpenAI and Microsoft in New York, brought by the New York Times and other publishers, Ziff Davis, and authors led by the Authors Guild. OpenAI is touting some 20 filings on its side, including a Justice Department statement of interest and amicus briefs from EFF, Public Knowledge, Authors Alliance, and CCIA, arguing that training AI models on copyrighted works is fair use. The most revealing fight is over “market dilution,” the theory Judge Vince Chhabria floated in Kadrey v. Meta last year: that even if an AI’s outputs infringe nothing, a flood of competing works can still count as market harm to the works it learned from.¹ Authors Alliance devoted its entire brief to asking the court to hold that “aggregate competitive pressure from noninfringing outputs” is not cognizable harm, and CCIA calls dilution “an unrecognized and unrealistic theory of harm.” They’re right, but notice what the theory concedes: the complaint is not that anyone copied your work, but that someone else produced new, non-infringing works that people might prefer—that is, competition. As I’ve said for years, IP advocates hate competition, and copyright, like patent, is at bottom a state-granted privilege to suppress it. If courts accept “dilution” as harm, copyright will no longer even pretend to protect expression; it will be an openly protectionist cartel for incumbent authors and publishers, and the only real fix is to abolish it.