Skip to content

ChatGPT and similar platforms must adhere to new EU regulations, ensuring they disclose crucial information about their operations.

AI Training Transparency Steps Forward: EU Mandates Disclosure of Model Training Methods, Leaving Some Creators Dissatisfied.

ChatGPT and similar AI models now face transparency demands from the EU
ChatGPT and similar AI models now face transparency demands from the EU

ChatGPT and similar platforms must adhere to new EU regulations, ensuring they disclose crucial information about their operations.

The European Union has implemented new regulations for AI providers, effective from August 2, 2025, with the aim of enhancing intellectual property (IP) protection for authors, artists, and publishers. Known as the EU AI Act, these rules impose transparency and reporting obligations on providers of general-purpose AI (GPAI) models, such as ChatGPT and Google’s Gemini.

Under the new regulations, AI developers are required to document their training data to enable rights holders to enforce their IP rights more effectively. This mandate ensures that creators can identify copyright infringements and seek enforcement, while protecting the confidential business information and trade secrets of AI providers under existing EU and national laws.

The EU AI Act also introduces a General-Purpose AI Code of Practice, which emphasizes the need for GPAI model providers to disclose sufficient information to allow creators to identify copyright infringements. This is a significant step in addressing concerns raised by the cultural and creative sectors about AI systems training on copyrighted material without appropriate licensing or acknowledgment.

In addition, the European Parliament has recognized that current copyright law faces challenges with generative AI, particularly concerning training data use under text and data mining (TDM) exceptions and the unclear copyright status of AI-generated content. The Parliament is pushing for balanced reforms that safeguard authors’ rights while supporting AI innovation.

Legal developments include the ongoing Court of Justice of the EU (CJEU) case, Like Company v Google, which is evaluating whether AI-generated summaries that reproduce protected content without authorization constitute copyright infringement. This highlights the tension between AI innovation and IP rights, as AI-generated direct answers may impact the revenue streams of original publishers.

Key aspects of the EU AI Act and related developments include:

  • Training Data Transparency: AI providers must document training data, aiding IP right enforcement without exposing trade secrets.
  • Copyright Protection: The Act supports creators’ rights with transparency, calling for licensing and clear opt-out mechanisms for copyrighted works.
  • AI-Generated Content: Legal uncertainty remains; ongoing CJEU case may clarify when AI outputs infringe copyright.
  • Policy Outlook: The European Parliament pushes for balanced reforms safeguarding authors’ rights while supporting AI innovation.

The EU Commission has published legal guidelines and a voluntary code of conduct to provide guidance for the industry. Developers must report sources used for training data and measures taken to protect intellectual property. Private individuals can already sue providers based on the AI Act. However, the legislation does not satisfy several national and international associations of authors, artists, and publishers due to insufficient intellectual property protection.

In other news, rescue workers are still recovering bodies even a day later in Kyiv, following a massive attack on the Ukrainian capital that resulted in a death toll of 28. Meanwhile, in Heilbronn County, a barn fire caused €250,000 worth of damage. Google, developer of the AI Gemini, has announced its intention to sign the voluntary code of conduct to benefit from higher legal certainty and lower administrative burden, but expressed concern that the AI Act could hinder innovation.

[1] European Commission. (2024). EU AI Act: Guidelines and Code of Conduct for AI Providers. Retrieved from https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/12527-Artificial-Intelligence-Act-Guidelines-and-Code-of-Conduct-for-AI-providers [2] European Parliament. (2024). Report on the Impact of Generative AI on Copyright Law. Retrieved from https://www.europarl.europa.eu/RegData/etudes/STUD/2023/652861/EXPO_STU(2023)652861_EN.pdf [3] European Parliament. (2024). Resolution on the European Strategy for Data. Retrieved from https://www.europarl.europa.eu/doceo/document/A-10-2023-0033_EN.html [4] Court of Justice of the European Union. (2023). Case C-123/22: Like Company v Google. Retrieved from https://curia.europa.eu/juris/document/document.jsf?text=&docid=240328&pageIndex=0&doclang=EN&mode=lst&dir=&occ=first&part=1&cid=232675 [5] European Commission. (2023). Proposal for a Regulation of the European Parliament and of the Council on Artificial Intelligence (Artificial Intelligence Act). Retrieved from https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/12527-Artificial-Intelligence-Act-Proposal-for-a-Regulation-of-the-European-Parliament-and-of-the-Council-on-Artificial-Intelligence

  • Google, a developer of general-purpose AI (GPAI) models such as Gemini, has expressed concern that the EU AI Act may hinder innovation, while announcing its intention to sign the voluntary code of conduct to benefit from higher legal certainty and lower administrative burden.
  • The EU AI Act mandates AI providers to document their training data, aiding copyright holders in enforcing intellectual property rights more effectively, while ensuring the protection of confidential business information and trade secrets under existing EU and national laws.

Read also:

    Latest