--- title: "AI and Intellectual Property: Training Data, Copyright and Disclosure" description: "How EU copyright law and the AI Act treat training data, rights holders' opt-outs, training summaries, and the protection of AI-generated output and inventions." url: https://www.pitch.law/academy/ai-intellectual-property-training-data-copyright-disclosure --- [pitch](https://www.pitch.law/) [Academy](https://www.pitch.law/academy) [Legal Aspects of Artificial Intelligence](https://www.pitch.law/academy/topics/legal-aspects-of-artificial-intelligence) # AI and Intellectual Property: Training Data, Copyright and Disclosure How EU copyright law and the AI Act treat training data, rights holders' opt-outs, training summaries, and the protection of AI-generated output and inventions. - By By Bart Lieben, Attorney-at-Law - Published 25 February 2026 ## The legal framework AI raises intellectual property questions at two points: the input, when models are trained on third-party material, and the output, when they generate text, images, code or designs. ### Training data: the text and data mining exceptions Directive (EU) 2019/790 on copyright in the Digital Single Market contains two exceptions for text and data mining (TDM), meaning automated analysis of text and data to generate information. - **Article 3** permits research organisations and cultural heritage institutions to reproduce and extract lawfully accessible works for scientific research. Rights holders cannot opt out. - **Article 4** permits anyone to reproduce and extract lawfully accessible works for TDM, including commercial purposes, but only if the rights holder has not reserved its rights. For content made available online, the reservation must be made in an appropriate manner, such as machine-readable means (Article 4(3)). Whether training a generative model falls within Article 4, and whether outputs that reproduce training material infringe, is not settled. A Hungarian court has referred the first questions on generative AI and copyright to the Court of Justice (Like Company v Google Ireland, C-250/25), concerning reproduction and communication to the public in the training and operation of a chatbot, and the press publishers' right. At the time of writing we have not verified a ruling. National courts have begun to decide cases, with differing approaches. Treat the position as developing. ### The AI Act: copyright policy and training summary Since 2 August 2025, providers of general-purpose AI models must put in place a policy to comply with EU copyright law, in particular to identify and respect reservations of rights under Article 4(3) of the Directive, using state-of-the-art technologies (Article 53(1)(c)). They must also publish a sufficiently detailed summary of the content used for training, following the format published by the AI Office (Article 53(1)(d)). A voluntary General-Purpose AI Code of Practice, published in July 2025, contains a copyright chapter. These rules apply to providers, wherever they train the model, if they place it on the EU market. Enforcement powers of the Commission towards general-purpose model providers start on 2 August 2026. ### The output: authorship In EU law, a work is protected if it is the author's own intellectual creation, reflecting free and creative choices (the Court of Justice in _Infopaq_, C-5/08, and _Painer_, C-145/10). That standard presupposes a human author. Content generated by an AI system with no meaningful human creative contribution is unlikely to be protected. Where a person makes real creative choices, for example in selecting, arranging or substantially editing the output, protection may attach to their contribution. Prompts alone are a contested basis. ### Inventions For patents, the European Patent Office's Legal Board of Appeal held in the DABUS case (J 8/20) that an inventor must be a natural person. AI-assisted inventions can be patented when a person contributed to the inventive concept and is named as the inventor. ### Trade secrets and confidentiality Entering confidential information into an AI tool can destroy its secrecy if the provider may retain or reuse it. See [database rights and trade secrets](https://www.pitch.law/academy/database-rights-and-trade-secrets). ## In practice A Kortrijk publisher of trade handbooks finds that an AI vendor's model reproduces passages from its books. It checks whether its website carries a machine-readable reservation of rights, and whether it has the tools to prove the date and form of the reservation. It then writes to the vendor, asks for the training summary and the copyright policy, and prepares a licence proposal. A Ghent design studio uses an image generator to produce brand visuals. Its client wants exclusive rights. The studio cannot promise copyright in the raw output. It documents the creative choices its designers made, adds substantial human editing, and agrees contract terms on ownership, exclusivity and indemnity with the client and the tool's provider. A software company wants to fine-tune a model on customer support transcripts. It checks the customer contracts, the GDPR basis and any third-party content in the transcripts before it starts. ## Frequently asked questions ### Can we stop AI companies from training on our website? You can reserve your rights in machine-readable form (for example in a robots protocol, metadata or terms), which makes the Article 4 exception unavailable for commercial TDM. It is not certain how all forms of reservation will be judged. ### Is AI-generated content copyrightable? Not by itself. Human creative contribution may be protected. Contracts should not assume more than the law gives. ### Do we have to disclose what we used to train our own model? If you are a provider of a general-purpose AI model on the EU market, yes: a public summary and a copyright policy. Other providers face different duties. ### Does using AI-generated content expose us to claims? If outputs reproduce protected material, yes. Ask vendors for IP indemnities and keep human review. ## How this connects Our service on [data and database rights protection](https://www.pitch.law/solutions/data-and-database-rights-protection) covers training-data governance, opt-out strategy and contract terms. [Our contract review](https://www.pitch.law/technologies/clause-library-risk-review) checks AI vendor terms, [the Document Studio](https://www.pitch.law/technologies/contract-studio) prepares licences and ownership clauses, and [our IP portfolio](https://www.pitch.law/technologies/portfolio-renewals) records your copyright works. Related: [AI clauses in vendor contracts](https://www.pitch.law/academy/ai-clauses-in-vendor-contracts). ## In short - Text and data mining exception and opt-out - Copyright policy and training summary duties for model providers - Whether AI output is protected - Open questions before the Court of Justice ## Related solutions - [Solutions](https://www.pitch.law/solutions/data-and-database-rights-protection) ### [Data and Database Rights Protection](https://www.pitch.law/solutions/data-and-database-rights-protection) [How to protect the data and databases you build, through database rights, trade secrets, copyright and the contracts that hold them together.](https://www.pitch.law/solutions/data-and-database-rights-protection) ## Technology we use - [Technology](https://www.pitch.law/technologies/clause-library-risk-review) ### [Contract Review](https://www.pitch.law/technologies/clause-library-risk-review) [Send us the contract the other side drafted and we check it clause by clause against our playbook, with a risk ranking an attorney has verified.](https://www.pitch.law/technologies/clause-library-risk-review) - [Technology](https://www.pitch.law/technologies/contract-studio) ### [Document Studio](https://www.pitch.law/technologies/contract-studio) [Our Document Studio drafts and analyses legal documents in a secure, on-premise environment, and keeps the documents behind your matters up to date as new data comes in and procedural steps fall due.](https://www.pitch.law/technologies/contract-studio) - [Technology](https://www.pitch.law/technologies/portfolio-renewals) ### [IP Portfolio, Renewals and Deadlines](https://www.pitch.law/technologies/portfolio-renewals) [Your trademarks, designs, domain names and patents in one view, with deadlines docketed, renewals batched and every computed date confirmed by a person.](https://www.pitch.law/technologies/portfolio-renewals) ## More in Legal Aspects of Artificial Intelligence - ### [AI Act Timeline: What Applies When](https://www.pitch.law/academy/ai-act-compliance-timeline) [The application dates of the AI Act as of October 2026, including the postponement of the high-risk rules adopted in 2026.](https://www.pitch.law/academy/ai-act-compliance-timeline) - ### [AI Act Meets GDPR: Where the Two Frameworks Overlap](https://www.pitch.law/academy/ai-act-meets-gdpr-where-frameworks-overlap) [How the AI Act and the GDPR apply side by side to the same AI system: roles, impact assessments, automated decisions, transparency, data quality and special categories.](https://www.pitch.law/academy/ai-act-meets-gdpr-where-frameworks-overlap) - ### [AI Act Risk Classification](https://www.pitch.law/academy/ai-act-risk-classification-explained) [How to tell whether an AI system is prohibited, high-risk, subject to transparency duties or free of specific obligations under the AI Act.](https://www.pitch.law/academy/ai-act-risk-classification-explained) - ### [AI Clauses in Vendor Contracts](https://www.pitch.law/academy/ai-clauses-in-vendor-contracts) [Which clauses to expect and negotiate when you buy or supply AI: AI Act allocation, data use, output ownership, IP indemnity, liability and change control.](https://www.pitch.law/academy/ai-clauses-in-vendor-contracts) - ### [AI in HR and Recruitment](https://www.pitch.law/academy/ai-in-hr-and-recruitment-under-the-ai-act) [Which AI tools in recruitment and people management are prohibited or high-risk under the AI Act, what employers must do, and how national labour law and the GDPR add to it.](https://www.pitch.law/academy/ai-in-hr-and-recruitment-under-the-ai-act) - ### [The AI Literacy Obligation](https://www.pitch.law/academy/ai-literacy-obligation-businesses-overlooking) [Article 4 of the AI Act has applied since February 2025 to every company using AI. What it requires, how the 2026 amendment changed it and what to document.](https://www.pitch.law/academy/ai-literacy-obligation-businesses-overlooking) ## Let’s talk Tell us what you’re working on. We’ll get the right person on it and reply within one business day. The first chat is free. [Schedule a meeting](https://www.pitch.law/schedule)