Truace
Subscribe

A legal study argues US and EU copyright rules leave AI training on creative works unsettled

Study title: AI AND COPYRIGHT: FROM A DOCTRINAL CRISIS TO A HYBRID MODEL OF COLLECTIVE LICENSING

A peer-reviewed legal study examines the clash between generative AI developers, who need large datasets for training, and rights holders, who want protection and fair payment. It compares US and EU law and concludes that neither fair use nor the TDM exception resolves the conflict.

Read the study abstract

The explosive development of generative artificial intelligence (hereinafter – AI) has triggered a fundamental crisis within intellectual property law. A central conflict has emerged between artificial intelligence developers, who require access to vast datasets for model training, and rights holders, who seek to protect their works and receive fair compensation for their use. The purpose of this article is to demonstrate the inadequacy of existing legal doctrines in addressing the use of copyrighted data for tr…

By the Truvace desk1 min read1 sourcePeer-reviewedTRV-2026-2248· cite & verify
Peer-reviewed study: Legal Horizons

In brief

The study matters because creators' compensation and legal certainty for AI developers both depend on which governance model prevails. Its proposed hybrid compulsory collective licensing model is a scholarly proposal, and the text gives no evidence that it has been adopted or tested in practice.

Main points

  1. The article analyzes US and EU law using formal-legal, comparative-legal, systemic and synthesis methods.
  2. It calls fair use flexible but unpredictable, and the TDM exception formalized but ineffective.
  3. Systemic analysis looks at challenges AI poses to authorship, originality and economic incentives.
  4. The author proposes a hybrid model of compulsory collective licensing as the most rational alternative.

The problem

Existing US fair use litigation and the EU TDM opt-out are argued to leave AI training on copyrighted works legally unstable and unable to deliver meaningful compensation to creators.

The rundown

The source frames generative AI's growth as triggering a fundamental crisis in intellectual property law. The central dispute pits developers needing vast training datasets against rights holders seeking to protect works and be compensated.

The methodology combines analysis of US and EU norms with a comparison of the two approaches' relative effectiveness. Copyright is treated as an integrated system whose core elements, authorship, originality and economic incentives, are challenged by AI.

The conclusion is that continued reliance on legacy doctrines produces perpetual instability. The proposed remedy is a hybrid architecture of compulsory collective licensing, presented as a hypothesis formulated by the author.

Sources

  1. Peer-reviewedLegal Horizons2025-08-29

The debate