Source: Tech Policy Press

United States District Judge Sidney Stein is considering whether using copyrighted works to train AI models constitutes fair use in multidistrict litigation involving OpenAI and Microsoft. The parties have requested decisions without a trial on that question and related claims; the source reports no ruling resolving the fair-use dispute. A dozen similar cases were consolidated in Manhattan in April 2025, and Stein accepted 56 amicus briefs on 6 November 2026. OpenAI and Microsoft argue that their models learn from copyrighted works rather than reproduce those works. The publishers allege that the training involves copying without permission and produces AI products that compete with the publishers’ work. Plaintiff supporters include publishing and entertainment organizations, while defense supporters include technology associations, libraries and digital-rights groups; other participants support neither party or declare no side.

Justin Hendrix presents two resources developed with Claude Code: a visualization of cited judicial precedents and a directory of the briefs with selected quotations. Using each brief’s declared position, the tracker classifies 27 briefs as supporting plaintiffs, 18 as supporting defendants, seven as supporting neither party and four as declaring no side. Other guides produce different counts because those guides classify the direction of the arguments instead of the positions expressly declared. The displayed precedent network covers decisions cited by at least eight of the 56 briefs; a citation can distinguish a decision rather than endorse its application. The resources underwent accuracy reviews but may contain errors, and the AI model’s individual decisions on ambiguous citations were logged without editorial review. The tracker lists replies as due on 20 November and states that no date has been set for oral argument.