Transparency is a core pillar, calling for detailed training-data records, ongoing template refinements, and potential use of watermarking to track data provenance.

The report supports exemptions for scientific research and education from training material use, but requires proper authorization and remuneration for commercialization.

Overall, the report pushes for licensing-driven governance of GenAI within EU copyright law, but warns the sectoral focus could miss broader reforms of the EU framework.

Axel Voss underscored that generative AI must operate under the rule of law, with creators receiving transparency, legal certainty, and fair compensation.

A sectoral press exemption aims to protect democracy, with stronger control by publishers and journalists over licensing and use.

Tech lobby CCIA contends the proposal could hinder innovation and digital competitiveness, arguing existing rules already balance rights-holders and AI development.

Industry groups delivered mixed reactions: some welcomed stronger creator rights and licensing to ensure compensation, while others stressed enforcement of current laws to avoid compliance costs and stifling innovation.

Under current rules, materials can be used for text and data mining unless an author reserves rights, a point repeatedly raised by industry voices.

News media content should be protected and fully compensated if diverted by AI, with safeguards for media diversity and pluralism.

The European Parliament has adopted recommendations to protect copyright works used in AI training, emphasizing transparency, fair remuneration, and rights holders’ ability to exclude their content from AI training.

The CCIA argues for enforcing existing rules rather than adding new regulatory frictions, favoring a results-driven approach.

Current practice often uses journalistic content without consent or remuneration, undermining journalists’ rights and creating licensing inequities that favor large media groups.