LexLint Beta

Law / European Union

European Union

AI law

The EU's AI transparency and output-labeling regime sits almost entirely in Article 50 of the AI Act (Regulation (EU) 2024/1689), which took effect on 2026-08-02 and requires disclosure of AI interaction, machine-readable marking of synthetic output, and labeling of deepfakes and AI-generated public-interest text. The Digital Omnibus on AI (Regulation (EU) 2026/1744, in force since 2026-07-27) left Article 50 largely untouched, adding only a four-month transition, to 2026-12-02, for the machine-readable marking sub-duty on generative systems already on the market before 2026-08-02. A narrower, platform-side duty sits in Digital Services Act Article 35(1)(k): very large online platforms and search engines must, as one item on an illustrative list of systemic-risk mitigation measures, consider making manipulated content that appreciably resembles real persons or events distinguishable through prominent markings. This research checked and excluded the Political Advertising Regulation (Regulation (EU) 2024/900) as a candidate, since its labeling duties concern ad sponsorship and targeting, not AI generation.

Scraping law

This document covers one dimension of EU scraping law, the text-and-data-mining exceptions in the Copyright in the Digital Single Market Directive, and does not yet characterise the rest. Article 3 gives research organisations and cultural heritage institutions an exception for mining works they lawfully access for scientific research, and Article 7(1) makes any contrary contractual provision unenforceable. Article 4 gives everyone else the same permission for lawfully accessible works, but only where the rightholder has not expressly reserved the use, and it is absent from the Article 7(1) list, so a contract can displace it. The database sui generis right in Directive 96/9/EC, the computer-misuse rules in Directive 2013/40/EU, and personal data under Regulation (EU) 2016/679 all bear on scraping in the EU and are unresearched here.

17 instruments named 4 researched in detail As of 2026-08-14

Instruments

Each one links to its LexLint note, which carries what it requires and what it flags on.

AI training data

AI Act, Article 53 (obligations for providers of general-purpose AI models)

Regulation (EU) 2024/1689, Article 53 In effect Effective 2025-08-02 official consolidated Official Journal text, EUR-Lex

Providers of general-purpose AI models must draw up and keep up to date the model's technical documentation for the AI Office and national competent authorities, make information and documentation available to providers who intend to integrate the model into their own AI systems, put in place a policy to comply with Union copyright law and in particular to identify and comply with a reservation of rights expressed under Article 4(3) of Directive (EU) 2019/790, and draw up and make publicly available a sufficiently detailed summary of the content used for training, following a template provided by the AI Office. Paragraph 2 exempts models released under a free and open-source licence from the first two duties only, and not at all where the model carries systemic risk, so the copyright policy and the training-content summary bind every provider of a general-purpose model. Article 113(b) applies Chapter V, which contains this Article, from 2 August 2025, a year ahead of the Regulation's general application date.

AI transparency

AI Act, Article 50 (transparency obligations for AI systems and synthetic content)

Regulation (EU) 2024/1689, Article 50 In effect Effective 2026-08-02 official consolidated Official Journal text, EUR-Lex

Providers of AI systems that interact directly with people must ensure users are informed they are dealing with an AI system unless obvious from context, and providers of generative AI must mark synthetic audio, image, video, or text output in a machine-readable, detectable format. Deployers of emotion-recognition or biometric-categorization systems must inform exposed individuals, and deployers of deepfakes or AI-generated public-interest text must disclose the artificial origin unless the content underwent human review with editorial responsibility. The Article reaches providers and deployers outside the EU whose system's output is used in the EU. A four-month transition, to 2026-12-02, applies only to the machine-readable marking sub-duty for generative systems already on the market before 2026-08-02, added by the Digital Omnibus on AI.

Digital Services Act, Article 35(1)(k) (systemic risk mitigation, synthetic media marking)

Regulation (EU) 2022/2065, Article 35(1)(k) In effect Effective 2023-08-25 official consolidated regulation text, EUR-Lex

As one of the listed, non-exhaustive systemic-risk mitigation measures a designated very large online platform or search engine may adopt, the provider should ensure that generated or manipulated content resembling real persons, objects, places, or events, and that would falsely appear authentic, is distinguishable through prominent markings when presented on the platform, alongside an easy-to-use flagging function. This is materially narrower than AI Act Article 50: it binds only Commission-designated VLOPs and VLOSEs as a risk-mitigation option, not every provider or publisher as a freestanding labeling mandate.

Copyright TDM

DSM Directive, Article 4 (text-and-data-mining exception and rights reservation)

Directive (EU) 2019/790, Article 4 In effect Effective 2021-06-07 official consolidated Official Journal text, EUR-Lex

Article 4 requires Member States to provide an exception or limitation for reproductions and extractions of lawfully accessible works for the purposes of text and data mining, and paragraph 2 permits retaining those copies for as long as the mining requires. Unlike the Article 3 exception, which is confined to research organisations and cultural heritage institutions acting for scientific research, Article 4 is limited by neither purpose nor actor, and recital 18 records that mining techniques are widely used by private and public entities alike, including for the development of new applications and technologies. Paragraph 3 makes the exception conditional on the rightholder not having expressly reserved the use in an appropriate manner, which the Article illustrates as machine-readable means for content made publicly available online, so a reservation removes the exception rather than merely signalling a preference. Article 7(1) renders contractual override unenforceable for Articles 3, 5 and 6 but does not list Article 4.

This summary covers the instruments LexLint has researched to a stage. Instruments named in the corpus but not yet researched are counted above and are not listed here. Every entry carries its own primary source on its note page. This is a research index, not legal advice.