A consolidated group of copyright cases before US District Judge Sidney Stein in Manhattan has placed the legality of artificial-intelligence training squarely before the federal judiciary.
OpenAI and Microsoft contend that training AI systems on copyrighted material is transformative because the systems learn statistical relationships rather than simply reproducing protected expression.
The New York Times and authors including John Grisham, Jonathan Franzen and George R.R. Martin argue that unauthorised training exploits copyrighted works and threatens the markets supporting professional journalism and creative writing. Both sides have sought summary judgment.
-
Tate Brothers Denied Bail in UK Extradition Fight
-
HURIWA Questions Handcuffing of Hospitalised Detainee
The central doctrine is fair use, particularly whether the training process is sufficiently transformative and what effect generative AI has on the existing or potential market for copyrighted material.
Earlier American decisions involving other AI companies have recognised transformative elements in AI training while also warning that market harm can alter the fair-use analysis.
The outcome could become one of the most influential copyright precedents of the generative-AI era. A broad ruling for AI developers could strengthen current training practices, while a ruling favouring rights holders could accelerate licensing regimes and significantly alter the cost of developing large AI models.
READ MORE: Reuters
