EU extends copyright consultation as it seeks evidence on AI training

The European Commission has extended a consultation on whether the EU’s copyright framework needs additional measures to meet technological and market developments, pushing its deadline from Nov. 3 to Nov. 10. The exercise focuses on the effect of technology, including artificial intelligence, on effective copyright protection.

The consultation is seeking submissions from rightsholders, generative-AI providers and others in the AI value chain, as well as online intermediaries, collective-management bodies, organisers of sports and live events, researchers, public authorities, consumer groups and non-governmental organisations. Its evidence-gathering covers AI training, piracy of live events, performers’ pay and copyright’s role in scientific research.

It follows an earlier Commission call for evidence on targeted copyright modernisation, which was open for contributions until June 25. That review is assessing the practical effect of the 2019 Copyright Directive in the Digital Single Market, including whether it facilitated the use of copyright-protected content in the digital environment, improved licensing and created a fairer marketplace.

The Commission has identified generative AI as a challenge for licensing and enforcing rights. The earlier call also examined remuneration for performers and producers when recorded music is played in the EU, alongside access to and reuse of works for research.

The 2019 directive created a text-and-data-mining exception for research organisations and cultural-heritage institutions, as well as a broader exception that can cover some commercial AI uses unless a rightsholder opts out. Neither applies where material was accessed unlawfully. For online material, reservations under the broader exception must be machine-readable.

The EU AI Act requires general-purpose AI-model providers to maintain copyright-compliance policies, including policies for identifying and observing such opt-outs. Yet legal questions surrounding training and generated outputs remain before courts. In Case C-250/25, Like Company, the Court of Justice of the European Union is considering whether chatbot outputs can infringe copyright or press-publisher rights, and whether training a large language model involves a copyright-relevant reproduction.

The Commission said the new consultation built on feedback from the earlier call and on previous studies and assessments. It remains open for contributions through Nov. 10.