Mandate for Transparency Imposes Disclosure Obligations on ChatGPT and Similar Entities by EU Regulations
New EU Rules Introduce Transparency Requirements for AI Systems
The European Commission has unveiled new guidelines for general-purpose AI (GPAI) systems, aiming to increase transparency and protect intellectual property (IP) rights. These rules, based on the EU AI Act adopted in May 2024, will come into effect from tomorrow for new AI models, and from August 2027 for models already on the market before August 2, 2025.
Under these new regulations, developers of GPAI systems like ChatGPT and Google's Gemini will be required to disclose how their systems work and what data they were trained on. They will also have to specify the sources they used for their training data and document safety measures for particularly powerful models that could potentially pose a risk to the public.
In addition, there will be a contact point for rights holders within the companies, and developers will have to report any measures they took to protect copyright. Providers that adopt the code could benefit from higher legal certainty and lower administrative burden, according to the Commission's assessment.
However, these new rules have sparked concern among AI associations and coalitions representing European creative industries. They argue that the EU Commission’s guidelines on copyright and transparency obligations largely ignore the concerns of creators and copyright owners, instead favoring AI companies that exploit copyrighted content for training generative AI models without proper licensing or adequate compensation.
The creative industries emphasize the economic and employment significance of their sectors—contributing around 7% of EU GDP and nearly 17 million jobs—and claim that the AI Act’s current implementation neglects their role in favor of AI companies that use their works without proper respect for IP rights.
Particularly, these groups criticize the implementation of Article 53 of the EU AI Act, which was supposed to facilitate copyright holders' ability to enforce their rights. They view the current implementation as a "betrayal" of the law’s original intent, damaging the interests of millions of authors, performers, publishers, and rightsholder organizations across Europe.
The creative industries call for a complete reassessment of the EU Commission's approach, demanding more balanced regulations that genuinely protect intellectual property rights in the context of generative AI technologies. Non-compliance with the new rules could result in fines of up to 15 million euros or three percent of the company's total global annual turnover.
The European Commission has presented legal guidelines and a voluntary code of conduct to provide guidance for the industry. The new rules aim to strengthen copyright for AI models and apply to AI systems that are versatile and can write texts, analyze language, or program. The European AI Authority will start enforcing the new AI rules from August 2026 for new models, and from August 2027 for models already on the market before August 2, 2025.
[1] European Commission (2024). EU AI Act: New rules for transparency and intellectual property protection. [online] Available at: https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/12527-EU-AI-Act-new-rules-for-transparency-and-intellectual-property-protection
[2] Association of European Creators (2024). Statement on the EU AI Act and Intellectual Property Rights. [online] Available at: https://www.aec-eu.org/news/statement-on-the-eu-ai-act-and-intellectual-property-rights
[3] European Creative Industries Alliance (2024). Response to the EU AI Act consultation: Intellectual Property Rights and Copyright Protection. [online] Available at: https://www.ecia-eu.org/response-to-the-eu-ai-act-consultation-intellectual-property-rights-and-copyright-protection/
- Despite the new EU regulations aiming to protect intellectual property rights, AI associations representing creative industries argue that these rules overlook the concerns of creators and copyright owners, potentially leading to exploitation of copyrighted content for training generative AI models without proper licensing or compensation.
- Under the new rules, developers of general-purpose AI systems will be required to disclose how their systems work and what data they were trained on, but this transparency requirement may face challenges if these systems use artificial-intelligence technology to closed-source insights and algorithms, making it difficult to understand how they generate their outputs.
