EUIPO released The Development Of Generative Artificial Intelligence From a Copyright Perspective

 In May 2025, the European Union Intellectual Property Office (EUIPO), commissioned by the University of Turin Faculty of Law and the Nexa Center for Internet & Society at the Polytechnic University of Turin, released the Study on the Development of Generative Artificial Intelligence from a Copyright Perspective. This report systematically analyzes the challenges and mitigation strategies of GenAI (Generative Artificial Intelligence) for copyright law across economic, technical, and legal dimensions. It emphasizes balancing stakeholder interests through an "​​exception clause + opt-out​​" mechanism, which does not outright invalidate the legality of AI training. Key findings include:
 
​​Training Phase​​:
 
GenAI training requires copying copyrighted content, potentially infringing reproduction rights. The Digital Single Market Copyright Directive (CDSM) provides a legal exemption for text and data mining (TDM) under strict conditions, preserving copyright holders' partial control over their works .
 
​​Opt-Out Framework​​:
 
A "​​default use + opt-out​​" approach balances innovation and protection. AI developers are permitted to use legally obtained content for TDM by default, reducing transaction costs. Copyright holders may declare prohibitions via standardized tools (e.g., metadata tagging, robots.txt protocols, machine-readable statements in terms of service) to exclude their works from TDM. However, existing opt-out mechanisms lack unified standards, leading to inconsistent enforcement. EUIPO announced plans to launch a ​​Copyright Knowledge Hub​​ by late 2025 to centralize TDM opt-out declarations and provide infringement screening tools .
 
​​Transparency Requirements​​:
 
To comply with the Artificial Intelligence Act (2024), AI-generated content must embed machine-readable identifiers (e.g., digital watermarks, fingerprints) for traceability. Current technical solutions remain insufficient, necessitating multi-technology integration .
 
​​Licensing Support for SMEs​​:
 
Public institutions should promote automated licensing platforms (e.g., smart contracts) to reduce SMEs' compliance costs and foster data licensing markets, ensuring transparent permission management and royalty distribution .
 
The report underscores that copyright law aims to incentivize innovation while safeguarding creators' rights: "​​Protecting the source of creativity ensures sustainable, high-quality data inputs for AI​​." Challenges persist, including non-EU entities circumventing opt-out declarations and cross-border enforcement uncertainties.
 
 
Original link：The Development of Generative Artificial Intelligence From a Copyright Perspective