Legislation inventory · Public Beta
This inventory is separate from case law. Bills and drafts are not enacted law. Statutory Background records predate the generative-AI wave and are labelled accordingly.
Regulation (EU) 2024/1689, as amended including Regulation (EU) 2026/1744
Requires GPAI model providers to maintain a policy to comply with Union copyright law, including identifying and respecting rights reservations under DSM Directive Art. 4(3), and to publish a sufficiently detailed summary of training content. Article 50 establishes transparency duties for synthetic/deepfake content.
View details →Directive (EU) 2019/790
Provides EU text-and-data-mining exceptions/limitations. Article 4 permits qualifying reproductions/extractions for TDM subject to lawful access and permits rightsholders to reserve their rights in an appropriate manner; for publicly available online content the reservation should be machine-readable.
View details →Legge 23 settembre 2025, n. 132
Amends Italian copyright law so protected works must result from human intellectual work even where created with AI assistance, and adds an express provision for reproductions/extractions through AI systems or models, including generative AI, subject to the existing TDM framework.
View details →Act No. 48 of 1970, Art. 30-4
Permits exploitation of works, to the extent considered necessary, where the purpose is not to enjoy the thoughts or sentiments expressed in the work, subject to statutory conditions and a proviso for unreasonable prejudice to the copyright owner. Japan Copyright Office guidance identifies AI development/training as a potential non-enjoyment use.
View details →Act No. 21311 (current consolidated text)
Requires advance notice when products/services are based on high-impact or generative AI, marking that generative-AI outputs were AI-generated, and clear disclosure/marking of realistic synthetic audio, images or video. Creative/artistic works may use disclosure methods that do not impede display or enjoyment.
View details →Order No. 15
Requires generative-AI service providers to respect intellectual property and use data/base models with lawful sources; where IP is involved training data must not infringe others rights. It also incorporates content-labeling obligations.
View details →国信办通字〔2025〕2号
Requires explicit and implicit labels for AI-generated/synthesized text, images, audio, video and virtual scenes, with duties for generation services, distribution platforms and users and restrictions on malicious removal or alteration of labels.
View details →2025 c. 18, ss. 135-137
Requires government economic-impact assessment and reporting on the use of copyright works in AI development, including TDM, technical controls, disclosure, licensing/remuneration and enforcement. The Government published the required Copyright and AI report and impact assessment in March 2026.
View details →H.R. 7209, 119th Congress
Would create an administrative subpoena process to help copyright owners determine which copyrighted works were used in training AI models.
View details →S. 3813, 119th Congress
Would require notices to the Register of Copyrights concerning copyrighted works used in building generative AI models.
View details →S. 1396, 119th Congress
Would establish content-provenance and authentication protections and restrict certain unauthorized uses/removal of protected provenance information, including in AI-related contexts.
View details →S. 4591 / H.R. 8915
Would create federal rights and remedies against unauthorized highly realistic digital replicas of an individuals voice or visual likeness, subject to defined exceptions and service-provider procedures.
View details →H.R. 9112, 119th Congress
Would give visual artists a proposed exclusive right concerning specified commercial exploitation or public distribution of stylistic impersonations.
View details →PL 2338/2023
The Senate-approved AI bill contains direct copyright provisions addressing protected content in AI development/training and related rightsholder and transparency issues. It was transmitted to the Chamber and remains under Special Committee consideration.
View details →RUU Hak Cipta (revision of Law No. 28/2014)
The DPR copyright reform contains a dedicated AI framework. The working text denies copyright to purely AI-generated output without human intellectual involvement, requires specified AI-use disclosures, addresses attribution of authorship in AI-assisted creation and establishes creator/personality protections.
View details →KUU Alm.del Bilag 232 (consultation draft)
Proposes copyright-law protections against realistic digitally generated imitations of performers/artists and a broader protection against non-consensual realistic digital imitation of personal characteristics such as appearance or voice.
View details →DOF 14-05-2026
Amends Mexican labor and copyright legislation to strengthen performer-worker protections in relation to uses of voice, image and likeness, including AI/digital replication contexts.
View details →Initiative of 30 September 2025
Would prohibit producers/employers from using specified voice, image, interpretive style, biometric attributes or artistic/labor materials to train, improve or feed commercial/productive AI systems without express, free, prior and informed consent, with written contractual terms including purpose, duration, remuneration and withdrawal.
View details →Initiative of 3 March 2026
Would recognize copyright in content generated by AI systems where there is significant human creative intervention and would expressly protect computer programs made by a natural person using AI.
View details →Legal Notice No. 88 of 2025, Rule 27
Requires responsible disclosure and human editorial review for AI-assisted media content, respect for intellectual property, and restrictions on cloning a real persons likeness, voice or persona without informed consent subject to exceptions.
View details →Senate Bills No. 4 of 2026
A proposed national AI governance framework that expressly defines synthetic media generated or manipulated using generative AI. Its public relevance to this map is limited to the synthetic-media / generative-AI dimension rather than copyright training.
View details →Promulgated 14 January 2026
Establishes transparency and explainability principles under which AI outputs should carry appropriate disclosures or labels, and requires government AI-data policy to protect intellectual property in training and output results.
View details →No. 230-VIII, 17 November 2025
Contains one of the most explicit AI/copyright statutory regimes identified: human creative contribution is required for copyright in AI-assisted works; qualifying prompts may be protected; training is excluded from specified educational/scientific free-use rules; rightsholders can express a machine-readable prohibition on training; and synthetic AI results are subject to marking/disclosure rules.
View details →Law No. 134/2025/QH15
Makes compliance with intellectual-property law an express duty in data collection/processing/use for AI development, training, testing and operation, and establishes marking/disclosure duties for AI-generated or modified content including synthetic likeness/voice and potentially misleading content.
View details →Copyright Act 2021, ss. 243-244
Singapore copyright law permits qualifying copying of lawfully accessed works for computational data analysis, expressly including uses such as text/data mining and training machine learning, subject to statutory safeguards.
View details →Royal Decree M/169 (published 13 February 2026)
Expressly permits specified copying of a lawfully published work for developing AI products and algorithms without authorization or compensation where the original copy was lawfully obtained and copying is limited to what the purpose requires, subject to the statutes general safeguard concerning normal exploitation and legitimate interests.
View details →Bill 2/3634
Would require licensing for appropriate remuneration for specified uses of protected works and related-rights subject matter in AI training, development, fine-tuning, testing, evaluation, dataset creation and model improvement, and would address certain commercial AI outputs that substitute for protected subject matter. The mechanism uses collective representation with an opt-out.
View details →Sénat No. 220 (2025-2026); Assemblée nationale No. 2634
Would establish a rebuttable presumption that protected copyright/related-right subject matter was exploited by an AI system where an indicator connected with the systems development, deployment or generated output makes such exploitation plausible.
View details →121/000096
Would establish Spains national AI governance/enforcement framework and makes failures to comply with specified AI Act transparency duties concerning machine-detectable synthetic content/deepfakes enforceable under national sanctions.
View details →Parliamentary Document 56K1540
Proposes stronger legal protection of natural persons digital likenesses, particularly against deepfakes and other synthetic manipulations.
View details →House Bill No. 9981
Would exclude purely AI-generated material from copyright; protect qualifying human expression/control in AI-assisted works; state that ordinary prompt engineering alone is insufficient absent substantial additional human creative contribution; and require disclosure of AI utilization and human/AI portions in registration.
View details →Boletín 17795-19
Would regulate realistic digital imitations generated by AI of a persons image, body or voice. The Chamber approved the proposal in general by 128-2 on 22 July 2026 and returned it to committee after amendments.
View details →Boletín 17618-19
Would require a clear and traceable identification mechanism for AI-generated content, aimed at transparency and mitigation of malicious synthetic-content uses.
View details →