#
Artificial Intelligence Act (Valid from 2024-08-01): Horizontal EU regulation setting risk-based rules for AI systems and general-purpose AI models, applying in stages from 2 February 2025 to 2 August 2028 as amended by Regulation (EU) 2026/1744.
6 instruments (6 current) · 8 obligations · 9 claims · 14 milestones
Instruments
- Artificial Intelligence Act Valid from 2024-08-01
Horizontal EU regulation setting risk-based rules for AI systems and general-purpose AI models, applying in stages from 2 February 2025 to 2 August 2028 as amended by Regulation (EU) 2026/1744.
- Code of Practice on transparency of AI-generated content Date not recorded
Voluntary code on marking, detection and labelling of AI-generated content, confirmed adequate by the Commission and AI Board.
- Commission guidelines on Article 50 transparency obligations Valid from 2026-07-20
Commission guidelines defining the scope of Article 50 transparency obligations, published 20 July 2026.
- Denmark: draft Copyright Act amendment on realistic digital imitations (deepfakes) Draft, no legal force
Draft Danish copyright amendment protecting people's likeness and voice against realistic AI imitations, with satire and parody carve-outs.
- Digital Omnibus on AI Valid from 2026-07-27
Amends the AI Act timeline, adds Article 5 prohibitions on non-consensual intimate deepfakes and CSAM-type material, replaces Article 4 and adjusts sandbox and AI Office powers.
- Italian Law on artificial intelligence Valid from 2025-10-10
Italian national framework law designating AgID and ACN, regulating AI in work and professions, copyright and criminal law.
Obligations
- Article 50(1)
Providers must design AI systems that interact directly with natural persons so that those persons are informed they are interacting with an AI system, unless obvious.
- Article 50(2)
Providers of AI systems, including general-purpose AI systems, generating synthetic audio, image, video or text must mark outputs in a machine-readable format and make them detectable as artificially generated or manipulated.
- Article 50(3)
Deployers of emotion recognition or biometric categorisation systems must inform the natural persons exposed to the operation of the system.
- Article 50(4)
Deployers of AI systems that generate or manipulate image, audio or video content constituting a deep fake must disclose that the content has been artificially generated or manipulated.
- Article 1(39)(b), inserting Article 111(4)
Providers of generative AI systems placed on the market before 2 August 2026 must take the necessary steps to comply with Article 50(2) by 2 December 2026.
- Art. 26(1)(c), inserting Art. 612-quater of the Penal Code
Unlawfully disseminating without consent images, video or voices falsified or altered with AI systems and capable of misleading about their genuineness, causing unjust harm, is an offence.
- § 5(1)
Traficom supervises, as market surveillance authority, compliance with the Article 50 transparency obligations of the AI Act.
- Draft Copyright Act amendment (TRIS 2025/0654/DK)
Unauthorised publication of realistic digital imitations of a person would be illegal and reportable to online sharing platforms as illegal content, with exceptions for satire and parody.
Milestones
- Entry into force of the AI Act Met
- Chapters I and II (definitions, prohibitions) apply Met
- Governance, GPAI obligations and penalties apply; Member States designate authorities Met
- Italian Law 132/2025 enters into force Met
- Omnibus enters into force Met
- General application; Article 50 transparency rules and enforcement start Met
- Article 5(1)(ba), (bb), (1a), (1b) prohibitions begin to apply Upcoming
- Article 50(2) compliance deadline for generative AI systems placed on the market before 2 August 2026 Upcoming
- Danish minister's hoped-for start of the deepfake copyright amendment Upcoming
- Providers of GPAI models placed on the market before 2 August 2025 must comply Upcoming
- National AI regulatory sandbox operational in each Member State Moved
- Chapter III Sections 1-3 apply to Annex III high-risk systems Moved
- Chapter III Sections 1-3 apply to Annex I (product-embedded) high-risk systems Moved
- Delegated acts adding AI health and safety requirements to the Machinery Regulation apply Upcoming
Claims and evidence (9)
Article 50 requires AI-interaction disclosure, machine-readable marking of synthetic content, emotion/biometric notice, and deep-fake disclosure, all applying from 2 August 2026.
Evidence and sources
- Regulation (EU) 2024/1689 (Artificial Intelligence Act), Official Journal text
Providers shall ensure that AI systems intended to interact directly with natural persons are designed and developed in such a way that the natural persons concerned are informed
- Regulation (EU) 2024/1689 (Artificial Intelligence Act), Official Journal text
Providers of AI systems, including general-purpose AI systems, generating synthetic audio, image, video or text content, shall ensure that the outputs are marked in a machine-readable format
- Regulation (EU) 2024/1689 (Artificial Intelligence Act), Official Journal text
Deployers of an emotion recognition system or a biometric categorisation system shall inform the natural persons exposed thereto of the operation of the system
- Regulation (EU) 2024/1689 (Artificial Intelligence Act), Official Journal text
Deployers of an AI system that generates or manipulates image, audio or video content constituting a deep fake, shall disclose that the content has been artificially generated or manipulated.
- Guidelines on transparency obligations for providers and deployers of AI systems
These guidelines define the scope of transparency obligations for providers and deployers of AI systems under Article 50 of the AI Act.
- Regulation (EU) 2024/1689 (Artificial Intelligence Act), Official Journal text
Providers of generative AI systems placed on the market before 2 August 2026 must comply with Article 50(2) by 2 December 2026.
Evidence and sources
- Regulation (EU) 2026/1744 (Digital Omnibus on AI), Official Journal text
Providers of AI systems, including general-purpose AI systems, generating synthetic audio, image, video or text content, that have been placed on the market before 2 August 2026 shall take the necessary steps in order to comply with Article 50(2) by 2 December 2026.';
- AI Act Service Desk - implementation timeline
02 Dec 2026 New prohibitions + Article 50(2) transition apply ** - Prohibitions, for AI systems generating non-consensual sexual deepfakes and child sexual abuse material, start to apply.
- Regulation (EU) 2026/1744 (Digital Omnibus on AI), Official Journal text
The Commission and the AI Board confirmed the Code of Practice on transparency of AI-generated content is an adequate voluntary tool, and more than 180 organisations have signed it.
Evidence and sources
- Code of Practice on transparency of AI-generated content
The Commission and the AI Board have confirmed that the code is an adequate voluntary tool to demonstrate compliance with the AI Act transparency obligations.
- Press release IP/26/1714: Commission starts enforcing AI Act rules and new transparency requirements on 2 August
180 organisations that have signed the Code of Practice on transparency of AI-generated content that operationalises the rules on transparency of AI- generated content.
- Code of Practice on transparency of AI-generated content
The Commission published guidelines on Article 50 transparency obligations on 20 July 2026.
Evidence and sources
- Guidelines on transparency obligations for providers and deployers of AI systems
These guidelines define the scope of transparency obligations for providers and deployers of AI systems under Article 50 of the AI Act.
- Guidelines on transparency obligations for providers and deployers of AI systems
Article 26 of Law 132/2025 inserts Article 612-quater into the Penal Code, criminalising unlawful dissemination of AI-generated or altered images, video or voices that cause unjust harm.
Evidence and sources
- Gazzetta Ufficiale Serie Generale n. 223 of 25-9-2025 (Legge 132/2025 full text)
Whoever causes unjust harm to a person by transferring, publishing or otherwise disseminating, without their consent, images, videos or voices falsified or altered through the use of artificial intelligence systems and capable of misleading as to their genuineness
- Gazzetta Ufficiale Serie Generale n. 223 of 25-9-2025 (Legge 132/2025 full text)
The Data Protection Ombudsman supervises Article 5 prohibited practices and Traficom supervises Article 50 transparency obligations in Finland.
Evidence and sources
- Laki eräiden tekoälyjärjestelmien valvonnasta 1377/2025
The Data Protection Ombudsman supervises, as market surveillance authority, compliance with the prohibited artificial intelligence practices under Article 5 of the AI Regulation.
- Laki eräiden tekoälyjärjestelmien valvonnasta 1377/2025
The Finnish Transport and Communications Agency supervises, as market surveillance authority, compliance with the transparency obligations under Article 50 of the AI Regulation.
- Laki eräiden tekoälyjärjestelmien valvonnasta 1377/2025
Denmark has notified to the European Commission, under TRIS procedure 2025/0654/DK, a draft act amending its Copyright Act to address realistic AI-generated digital imitations (deepfakes).
Evidence and sources
- Denmark's proposed copyright rules on deepfakes
Denmark has notified a draft act amending its Copyright Act to address realistic digitally generated imitations, commonly known as 'deepfakes'.
- Denmark's proposed copyright rules on deepfakes
Derived: Probable Denmark was due to introduce the deepfake bill to the Folketing on Thursday, 8 October 2026, with the minister hoping it will take effect at the start of 2027.
Evidence and sources
- Denmark introduces bill to protect people's faces and voices from deepfakes
Denmark is due to introduce a bill on Thursday that would ban the sharing of realistic deepfakes of people's faces and voices without their consent.
- Denmark introduces bill to protect people's faces and voices from deepfakes
The minister hopes the law will take effect at the start of 2027.
- Denmark introduces bill to protect people's faces and voices from deepfakes
Under the notified Danish draft, unauthorised publication of realistic AI-generated imitations of a person would be illegal and could be reported to online sharing platforms as illegal content, subject to satire and parody exceptions.
Evidence and sources
- Denmark's proposed copyright rules on deepfakes
As set out in the TRIS notification, such unauthorised publication would be illegal under Danish law, and the content could be reported to online sharing platforms as illegal content.
- Denmark's proposed copyright rules on deepfakes