GEMA v. OpenAI (Munich I Regional Court, 2025): doctrinal, comparative, and international perspectives on copyright and generative AI

This article analyzes the Regional Court of Munich I's decision in the GEMA v. OpenAI case, examining whether training generative AI models on copyrighted works constitutes infringement. The Court held OpenAI liable, finding that AI training creates copyright-relevant reproductions inside g...

Descripción completa

Detalles Bibliográficos
Autores: Dornis, Tim W., Ginsburg, Jane C., Lucchi, Nicola
Tipo de recurso: artículo
Fecha de publicación:2026
País:España
Institución:Universitat Pompeu Fabra
Repositorio:Repositorio Digital de la UPF
OAI Identifier:oai:dnet:rdupf_______::73195305ff55a765a2cf99794eb270ad
Acceso en línea:https://hdl.handle.net/10230/72984
http://dx.doi.org/10.1628/jz-2026-0084
Access Level:acceso abierto
Palabra clave:Intel·ligència artificial -- Dret i legislació
Propietat intel·lectual
Mineria de dades -- Dret i legislació
Descripción
Sumario:This article analyzes the Regional Court of Munich I's decision in the GEMA v. OpenAI case, examining whether training generative AI models on copyrighted works constitutes infringement. The Court held OpenAI liable, finding that AI training creates copyright-relevant reproductions inside generative models not sheltered by text and data mining exceptions under EU and German law. The authors explore the ruling's implications, compare current US fair use litigation, and assess both against the background of international copyright law, particularly the Berne Convention's three-step test.