EU Parliament Moves to Rein in AI Training on Copyrighted Content

Key Takeaways

• Parliament Demands Transparency: AI providers would need to disclose copyrighted works used in training datasets and maintain records of crawling and data collection.
• Creators Must be Compensated: Lawmakers insist that copyrighted material used to train generative AI systems should be subject to fair remuneration.
• New Licensing Framework Proposed: The European Commission is being urged to develop sector-based licensing markets for AI training data.
• News Media Protections Emphasized: MEPs want compensation and opt-out rights for publishers whose journalism is used by AI systems.
• AI-Generated Content not Eligible for Copyright: Parliament reaffirmed that works created entirely by AI should not receive copyright protection.

Deep Dive

The European Parliament has voted overwhelmingly to strengthen protections for copyrighted works used in artificial intelligence systems, showing a growing concern among lawmakers that generative AI is reshaping the economics of creative industries without clear rules for compensation or consent.

Members of the European Parliament (MEPs) adopted a set of recommendations Tuesday aimed at ensuring transparency, fair remuneration, and stronger safeguards for rights holders when their work is used to train AI systems. The proposal passed by a vote of 460 to 71, with 88 abstentions, reflecting broad political support across Parliament for tightening oversight of how generative AI interacts with copyrighted material.

At the center of the proposal is a push to clarify that EU copyright law applies to all generative AI systems placed on the EU market, regardless of where those systems were trained.

For Europe’s creative industries, an economic sector that lawmakers say accounts for 6.9 percent of the EU’s GDP, the vote reflects mounting concern that AI models are being trained on vast volumes of copyrighted works without clear compensation or permission.

Transparency and Payment for Training Data

MEPs argue that companies developing generative AI systems should be required to clearly disclose the copyrighted material used to train their models and compensate creators when their work contributes to the system’s capabilities.

Under the recommendations, AI providers and deployers would need to maintain detailed records of crawling activity and training data sources, including itemized lists of copyrighted works used in training datasets.

Failure to provide such transparency could be treated as potential copyright infringement, opening the door to legal challenges from rightsholders. If courts ultimately rule in favor of creators, AI providers or deployers would be responsible for covering all legal costs and associated expenses.

Lawmakers also asked the European Commission to explore mechanisms for ensuring remuneration for past uses of copyrighted material, though they rejected the idea of allowing AI companies to rely on a single flat-rate global licensing arrangement to cover training activity.

Another major element of the proposal calls on the European Commission to establish a new licensing framework for copyrighted works used in AI development.

The framework could include voluntary collective licensing arrangements organized by sector, allowing creators, publishers, and small and medium-sized enterprises to negotiate compensation for the use of their works in AI training.

MEPs also stressed that rightsholders must retain the ability to opt out of having their content used in AI training. To support that process, lawmakers suggested that the European Union Intellectual Property Office (EUIPO) could manage an official opt-out registry for creators who do not want their works used by AI developers.

Concerns Over News Media and Information Diversity

The Parliament’s recommendations place particular emphasis on protecting the news media sector, which lawmakers say faces unique risks as AI systems increasingly scrape, summarize, and redistribute journalistic content.

MEPs warned that AI tools that aggregate news content could divert traffic and advertising revenue away from media outlets while using their reporting to generate automated responses.

To address this, the proposal states that news organizations should be fully compensated when their content contributes to AI-generated outputs, and they should retain the right to refuse the use of their material for training AI systems.

Lawmakers also highlighted concerns about media pluralism, warning that AI aggregation tools could selectively process information or favor content from platforms that also operate AI services.

No Copyright Protection for AI-Generated Content

The Parliament’s recommendations also take a clear stance on the legal status of AI-created material. MEPs said content generated entirely by AI systems should not receive copyright protection, reinforcing the principle that copyright applies to works created by human authors.

They also called for stronger protections for individuals against the dissemination of manipulated or AI-generated content, emphasizing that digital service providers have a responsibility to act against illegal uses of such material.

Rapporteur Axel Voss, a German member of the European People’s Party (EPP), said the proposals are intended to strike a balance between encouraging AI innovation and protecting creators.

“We need clear rules for the use of copyright-protected content for AI training,” Voss said following the vote. “Legal certainty would let AI developers know which content can be used and how licences can be obtained. On the other hand, rightsholders would be protected against unauthorised use of their content and receive remuneration.”

“If we want to promote and develop AI in Europe while also protecting our creators, then these provisions are absolutely indispensable,” he added.

The recommendations now place pressure on the European Commission to translate the Parliament’s position into legislative or regulatory proposals, as policymakers across Europe continue to grapple with the rapidly evolving relationship between artificial intelligence and intellectual property rights.

The GRC Report is your premier destination for the latest in governance, risk, and compliance news. As your reliable source for comprehensive coverage, we ensure you stay informed and ready to navigate the dynamic landscape of GRC. Beyond being a news source, the GRC Report represents a thriving community of professionals who, like you, are dedicated to GRC excellence. Explore our insightful articles and breaking news, and actively participate in the conversation to enhance your GRC journey.