How ‘learnrights’ would compensate creators for AI model training
How ‘learnrights’ would compensate creators for AI model training
Publish Date: 2026-05-12 09:33:00
Source Domain: mitsloan.mit.edu
-
Copyright Law and Generative AI: The debate over whether generative AI can legally learn from copyrighted content without compensation has emerged, with significant implications for content creators. No binding determination on this issue has been made, and lawsuits, such as Encyclopedia Britannica’s against Open AI, highlight the conflict.
-
Learnright Laws: In response to these challenges, MIT’s Thomas Malone proposed “learnright” laws, which would grant copyright holders the exclusive right to license their content for AI model training. This idea aims to address how AI technologies impact incentives for human creativity.
-
Fair Use vs. Copyright: Current U.S. copyright law allows “fair use” for individuals to learn from and create new content based on copyrighted material, but generative AI can produce outputs dependent on massive, often non-direct, use of copyrighted content, raising new legal questions.
-
Legal Ambiguities and Uncertainties: Despite extensive lawsuits and legal challenges, the question of how to fairly compensate content creators for AI training has not been clearly resolved. No case has reached the Supreme Court, leaving the future of creative livelihood in an AI world uncertain.
-
Proposed Learnright Mechanism: The learnright approach involves an opt-in program where creators would register their copyrighted content, and AI companies could legally use it only with proper licensing and compensation negotiated directly with creators or their agents.
-
Benefits of Learnright: Malone and his co-authors argue that learnright laws would ensure that AI training does not undercut creator incentives, maintain ethical acknowledgment of original works, and ultimately create a more balanced and flourishing information marketplace benefiting both creators and AI providers.
