Source: PwC TLS Blog

The European Parliament resolution of 10 March 2026 recognises that generative artificial intelligence challenges the assumptions underpinning EU copyright law, in particular the text and data mining (TDM) exception established by Directive (EU) 2019/790. The current TDM framework permits large-scale extraction and analysis of content, including copyrighted works, allowing rights-holders to exercise a machine‑readable opt-out; the resolution highlights that generative models use such data not only for analysis but to train systems that can produce new, potentially substitutive outputs that affect the economic value of original works.

The resolution advocates shifting the TDM landscape from a broad, largely free exception toward arrangements closer to authorisation and remuneration, emphasising transparency and traceability. It notes practical limits of the opt-out—lack of uniform technical standards, scattered reservations across platforms, and verification difficulties—and proposes obligations on model providers to disclose detailed information on used content, with a presumption of copyright infringement if such information is unavailable. The text further flags risks at deployment (not only training), suggests creating a European licensing market and an EUIPO‑managed central registry for rights reservations, and stresses that while the resolution is non‑binding it signals the likely direction of EU policy.