Sittings · Document
Follow-up to the European Parliament non-legislative resolution on copyright and generative artificial intelligence – opportunities and challenges
Rapporteur: Axel VOSS (EPP / DE)
References: 2025/2058(INI) / A10-0019/2026 / P10_TA(2026)0066
Date of adoption of the resolution: 10 March 2026
Competent Parliamentary Committee: Committee on Legal Affairs (JURI)
Brief analysis/ assessment of the resolution and requests made in it:
The Commission welcomes the European Parliament’s resolution on opportunities and challenges in copyright and generative artificial intelligence (“AI”). In this resolution, the European Parliament considers that the existing EU framework, including the Regulation (EU) 2024/1689 (AI Act) and the Directive on copyright in the Digital Single Market (Directive EU 2019/790, DSM Directive), is insufficient to address the challenges of licensing copyright protected material for Generative AI (GenAI) and potential infringements of current copyright rules. It calls for the full enforcement of EU law and measures to facilitate voluntary licensing, strengthen transparency and ensure fair remuneration of rightsholders. The resolution underlines the specific challenges faced by the press and news media sector in terms of control over the use of their content by generative AI as well as the implication in terms of access to and diversity of information.
Response to requests and overview of actions taken, or intended to be taken, by the Commission:
EU copyright legal framework for copyright and GenAI (paragraphs 2, 4, 15)
The DSM Directive updated the EU copyright framework to take into account changes brought by emerging digital technologies. Its aim was to uphold a high standard of protection for human creativity and to ensure the adequate distribution of economic incentives across the value chain. The Commission therefore fully shares the Parliament's objective to ensure that the EU copyright framework maintains relevance in the context of the rapid development of generative AI and adequately addresses the challenges faced by rightsholders.
Moreover, the DSM Directive sought to establish a balanced approach that fosters innovation and stimulates investments in digital technologies, including economic activities that leverage digital content and data in innovative ways. Specifically, the DSM Directive introduced two new mandatory exceptions or limitations for text and data mining (TDM) to increase legal certainty and support research and innovation. The exception or limitation in Article 4 applies on condition that the use of works and other subject matter has not been expressly reserved by the rightsholders in an appropriate manner, such as machine-readable means in the case of content made publicly available online.
As indicated in recital 105 of the AI Act, where the rights to opt out have been expressly reserved in an appropriate manner, providers of general-purpose AI models need to obtain an authorisation from rightsholders if they want to carry out text and data mining over such works. In due course, the Court of Justice will rule on the interpretation of the exception in Article 4 of the DSM Directive in a case currently pending before the Court of Justice (C-250/25).
Under Article 53(1)(c) of the AI Act, providers placing General-Purpose AI (GPAI) models on the EU market must put in place a policy to comply with Union Law on copyright and related rights, and in particular to identify and comply, including through state-of-the-art technologies, a reservation of rights expressed pursuant to Article 4(3) of the DSM Directive. The obligations under Article 53(1) (c) and (d) are addressed to GPAI model providers and should support compliance with EU copyright law. These obligations regarding GPAI models are supervised and enforced by the AI Office.
The Commission is currently preparing the review of the DSM Directive to assess the effectiveness of the existing rules. This includes carefully examining the impact of the TDM exceptions and their relevance in the current environment, taking into account relevant market and technological developments.
The Commission will also explore the need for adjustments of the existing copyright acquis or for additional measures, to address the challenges faced by creators and creative industries in the context of AI.
Specific challenges of the press and new media sector (paragraphs 6, 7, 8)
The DSM Directive has introduced a new right for press publishers covering the online use of their press publications by information society service providers. The Commission will examine to what extent this right has helped press publishers to license the use of their content in the online environment and obtain appropriate remuneration.
The Commission will also evaluate the suitability of the current legislation to mitigate the adverse substitutive effects of various AI techniques such as retrieval-augmented generation (RAG) to preserve media pluralism and diversity of information. Regarding the phenomenon described in paragraph 6 of the resolution, the Commission notes that relevant questions concerning the rights of press publishers in the context of AI are the subject of a recent request for preliminary ruling (C-250/25) pending before the European Court of Justice.
Under the Digital Markets Act, gatekeepers must not treat their own services, including AI, more favourably in ranking than similar services offered by third parties. The EU is committed to fully embracing the digital and AI revolution and actively monitors market developments, including the ones concerning AI services. Ongoing efforts on this front include continuous monitoring and regulatory dialogue. If necessary, the Commission will employ all available tools to address them effectively.
Effective rights reservations and AI training (paragraph 10)
The availability of effective means for expressing rights reservations is a crucial element of the balanced policy approach pursued with the TDM exception in Article 4 of the DSM Directive. The interpretation of the meaning and scope of rights reservation under Article 4(3) of the Directive (EU) 2019/790 (DSM Directive) will ultimately be a matter for the Court of Justice, in light of the aims and objectives of the DSM Directive and the Union copyright acquis.
The Commission remains committed to supporting the practical application of the rights reservation mechanism, including in the context of the obligations under Article 53 of the AI Act. Signatories to the GPAI Code of Practice have committed to employ web crawlers that respect robots.txt and any subsequent version of this Protocol for which the Internet Engineering Task Force (IETF) demonstrates that it is technically feasible and implementable by AI providers and content providers, including rightsholders. Signatories to the Code also commit to comply with other appropriate machine-readable protocols for expressing rights reservations, either adopted by international and European standardisation organisations or otherwise identified in a process facilitated at EU level. The Commission is currently facilitating a process at EU level with the involvement of relevant stakeholders to identify and generally agree on state-of-the-art machine-readable solutions in the context of the AI Act implementation and the GPAI Code of Practice. The AI Office is supervising the compliance with Article 53 in the AI Act for all providers of GPAI models, regardless of whether they are signatories to the Code or not.
Taking stock of the challenges and limitations observed in relation to existing technical protocols and other rights reservation methods, the Commission is in the process of finalising a study investigating the opportunity and technical feasibility of introducing a registry of right reservations. Such registry would leverage state-of-the-art content-based digital fingerprinting technologies as a complementary means for rightsholders to express their rights reservations – including from GenAI. With this study, the Commission seeks to explore a solution that balances the interests of rightsholders and AI developers, ensuring that the rights of content creators are respected while facilitating the growth and innovation of AI technologies in Europe and ensuring there is no circumvention of the Union’s obligations under the Berne Convention by which it is bound under the Agreement on Trade-Related Aspects of Intellectual Property Rights and the World Intellectual Property Organization (WIPO) Copyright Treaty.
Supporting licensing, remuneration of rightsholders, and enforcement (paragraphs 9, 11, 21)
The resolution emphasises the necessity of facilitating voluntary licensing in order to enable rightsholders to generate revenue from the use of their protected content for GenAI applications. The Commission acknowledges the various licensing practices, including individual and collective licensing, and the distinct needs of different rightsholders across the creative sectors. The Commission will explore the need for new measures to facilitate the conclusion of licences for uses of copyright-protected content which ensure that various rightsholders can be remunerated in a way which respects the work and protected subject matter including its market value, with the objective to support legitimate access to high-quality content for AI providers. When doing so, the Commission will take into account the existing Union copyright acquis including the relevant provisions of the DSM Directive.
The Commission will also investigate measures to strengthen the enforcement position of rightsholders in relation to misuses of their protected works or other subject matter in the training of GenAI and its subsequent applications.
Measures against infringements of copyright in generative AI outputs (paragraph 26)
As announced in the Apply AI Strategy, the Commission will launch a targeted study on AI-generated content and on how technologies, including AI itself, can help prevent, detect and remove copyright-infringing contents, without affecting legitimate uses of copyright-protected content. The results of this study will contribute to the implementation of measures set out in the Copyright Chapter of the above-mentioned GPAI Code of Practice, which aims to mitigate the risk that a downstream AI system into which a GPAI model is integrated generates outputs that may infringe rights in works or other subject matter protected under Union law on copyright or related rights.
The Code of Practice is conceived as a dynamic and flexible instrument, designed to evolve alongside technological, market and regulatory developments. As a voluntary compliance tool supporting the implementation of the AI Act, it is intended to be updated in light of practical experience, stakeholder input and emerging challenges. This adaptive nature allows the Code to progressively incorporate more refined technical solutions, operational practices and standards, ensuring that it remains effective and responsive in a rapidly evolving AI ecosystem while complementing the broader EU legal framework.
Increased transparency on the use of copyright-protected content (paragraphs 12, 14, 24)
In accordance with Article 53(1)(d) AI Act, the Commission published a Template for the Public Summary of Training Content for GPAI models in July 2025. This template provides a clear and uniform framework for GPAI model providers to give rightsholders and the public insight into the data used to train their models, including their crawling practices. The transparency obligation concerning training content, while taking due account of the protection of trade secrets and confidential business information of GPAI model providers, aims to enable rightsholders to assess the use of their content and exercise their rights in an informed manner.
The AI Office, which will have effective enforcement powers as from August 2026, will monitor and assess the conformity of providers’ training data summaries with the template. The information provided through the template summary will specifically allow rightsholders to better assess what data modalities and type of content was used and facilitate parties with legitimate interests, including rightsholders, to exercise their rights under Union law.
The summaries of training content should include a description of public and private datasets, a list of all large publicly available datasets, and information regarding the data scraped from online sources. This includes names of crawlers used, period of collection, comprehensive description of the type of content crawled, and a list of the top 10% of all domains that have been scraped from the internet (for SMEs top 5% or 1000, whichever is lower).
As mentioned in the explanatory notice accompanying the Template, the Commission further recommends GPAI model providers to voluntary enable rightsholders to obtain additional information, upon request, on the use of their protected works which are available on specific internet domains. This mechanism does not affect other available remedies for rightsholders under Union law on enforcement of intellectual property rights (e.g. Article 8 of the Intellectual Property Rights Enforcement Directive).
As stated in the explanatory notice, the Commission will monitor the implementation of the Template and, where necessary, review the Notice and the Template, in view of practical experience gained and of the pace of technological, societal and market developments in this area.
On this basis, the Commission will explore the need for complementary mechanisms to help rightsholders obtaining the necessary information to effectively licence and enforce their rights.
Content labelling and protection against AI-generated imitations of personal characteristics (paragraphs 27 and 28)
As regards paragraph 28, Article 50 of the AI Act subjects AI-generated or-manipulated content to transparency obligations in order to support distinguishing AI-content from human content and minimise the risk of manipulation, deception and misinformation. Two obligations are particularly relevant for the creative and cultural sectors. First, providers of generative AI systems are obliged to mark the AI outputs (audio, image, video and text) in a machine-readable format and ensure they are detectable as artificially generated or manipulated. Such marking has to be effective, interoperable, robust and reliable as far as this is technically feasible, taking into account the specificities and limitations of various types of content, the costs of implementation and the generally acknowledged state of the art, as may be reflected in relevant technical standards. Second, deployers of generative AI systems (such as professional musicians or visual artists) that generate or manipulate image, audio or video content constituting a deep fake must disclose that the content has been artificially generated or manipulated. If the deep fake content forms part of an evidently artistic or creative work, disclosure can happen in an appropriate manner not hampering the display or enjoyment of the work.
To operationalize the marking and labelling obligations from Article 50 AI Act, the AI Office is currently facilitating the development of a stakeholder-driven Code of Practice at Union level, involving representatives from the creative and cultural sectors. The Code of Practice is a voluntary tool on which providers and deployers of generative AI systems can rely to comply in a practical way with the legal requirements. A second draft of the Code was published on 5 March 2026, and the final version is expected to be published in early June 2026. In parallel, the AI Office will also issue guidelines to provide further guidance for providers and deployers on the scope of the transparency obligations which will become applicable in August 2026.
With regard to paragraph 27, the Commission appreciates the European Parliament call to explore measures to protect individuals from AI-generated content that imitates their personal characteristics. The Commission will examine the issue within the context of copyright, personality rights or the AI policy frameworks and explore solutions to curb the deceptive use of deepfake technologies and ensure that performers are adequately protected.
To complement the above-mentioned actions and initiatives and as part of the implementation of the Culture Compass, a dedicated AI strategy for cultural and creative sectors will be developed with the aim to ensure that AI enables and reinforces human creativity, while safeguarding European cultural and linguistic diversity.
***
The Commission looks forward to continuing its close collaboration with the European Parliament to achieve these crucial objectives, ensuring the EU copyright acquis remains a relevant driver of economic, cultural and societal benefit in the age of AI.