The much-anticipated final trilogue on the AI Act has intensified discussions on the intersection of AI and copyright, particularly the Text and Data Mining (TDM) exception.

Originating from amendments introduced by the European Parliament (as part of transparency obligations for providers of foundation models to make publicly available a summary of the copyright-protected content used to train the AI algorithms), debates emerged around the topic with organisations representing rightholders supporting disclosure and others questioning its practicality.

The concept of foundation models, embraced by the European Parliament, appeared to have reached a consensus by the co-legislators in the latest trilogue held in October.

However, concerns raised from large Member States such as France, Germany, and Italy in a subsequent technical meeting led to a dispute.

The European Commission's intervention proposes a compromise, marking the first official reference to reinforcing the TDM exception, as laid down in the recent Copyright Directive.

Simultaneously, the TDM exception itself faces challenges in effective enforcement due to the lack of harmonised standards.

And so poses the question as to whether the existing copyright regime is suitable for the AI age.

Key figures in the debate advocate for a copyright revision signalling potential developments in the upcoming mandates of the European Commission and Parliament.

See the full article here.