Why is USA Today suing OpenAI over training data?

Numbers match the citations in the article. A working link isn't proof that a page supports a claim; check the quoted passage and date.

1. 1

   Umjetna inteligencija i mediji: The New York Times Company protiv kompanija OpenAI i Microsoft

   Research paper (Mešević)Published Oct 1, 2024Checked Oct 9, 2026

   “In addition, it addresses the possible safe harbours for AI-developers in this context. Firstly, in the European Union, in terms of the text and data mining exception from Directive (EU) 2019/790, which is strongly backed by the text of the EU AI Regulation. Secondly, in the United States of America, in terms of applying the legal doctrine of „fair use“, especially in the sense of the transformative use of works. The author elaborates in detail the circumstances of the origin and the content of the dispute and looks at parallel cases in the domain of relationship between copyright and artificial intelligence, which could provide some guidance regarding the direction of the outcome of the NYTC– case. In addition, the paper draws attention to positive examples of OpenAI concluding agreements and even creating partnerships with large media organizations (e.g. Associated Press, Le Monde or News Corp) regarding the use of their content in GPT-model training, as well as in the context of technological cooperation.”

2. 2

   The Lawfulness of Using Inventions for Generative AI Training : A Case Study of a US Lawsuit against OpenAI and Perplexity AI

   JUSTISI (Ismantara & Silalahi)Published Dec 25, 2024Checked Oct 9, 2026

   “GAI outputs may also infringe copyright if: (1) the source is not cited, violating Article 7 on copyright management information; (2) substantial portions of the work are reproduced, violating the rightholder's economic rights under Article 9; or (3) the work is distorted in a way that harms the rightholder’s honor, infringing on moral rights under Article 5. To accommodate AI development, specific regulations integrating AI transparency principles outlined in SE Kominfo 9/2023 are required. These regulations could include obligations for AI companies to release summaries of training datasets, include Uni EropaLAs that define the responsibilities of AI developers and users, and provide disclaimers regarding AI's limitations. Regarding the fulfillment of rightholders’ economic rights, a non-exclusive blanket license through Collective Management Organizations (CMOs) as stipulated in Permenkumham 15/2024 is necessary. These regulations should be synchronized with related policies to establish legal certainty that adapts to technological advancements.”

3. 3

   Strength in Numbers: Group Copyright Registration of Online News Articles as a Weapon in the Battle Against AI Infringement

   Utah law review (Farmer)Published Jan 1, 2026Checked Oct 9, 2026

   “In this fight, copyright protection remains an essential weapon for publishers to safeguard their content. However, U.S. copyright law has lagged behind the news industry’s integration of new technologies and mediums, with current procedures for staying abreast of copyright registration burdensome or infeasible. A new regulation from the U.S. Copyright Office, “Group Registrations of Updates to a News Website,” or 89 FR 311, allows a digital news publisher to collectively register a group of updates to a news website by submitting a representative portion of the content, rather than the complete contents of the website. It has the potential to resolve challenges standing in the way of online news publishers seeking this protection for their pages. I examine how published content is used by generative AI, current copyright protections, and how 89 FR 311 could strengthen copyright protections for news publishers while significant legal issues regarding generative artificial intelligence and the media await resolution.”

4. 4

   Assessing the feasibility of collective licensing of in-copyright works as training data for generative AI systems.

   Proceedings of the National Academy of Sciences of the United States of America (Samuelson)Published Jul 20, 2026Checked Oct 9, 2026

   “If fair use defenses succeed, developers will be free to continue to commercially exploit models already built on copyrighted data as well as to use these data to train new models or fine-tune existing ones. If copyright owners prevail, developers may be liable for billions of dollars of damages. Developers could also be enjoined from further model development on in-copyright works and even ordered to destroy models trained on infringing works. Numerous commentators have proposed collective licensing as a compromise solution to the copyright-training-data dilemma. Other commentators have questioned the feasibility of such a compromise. This article discusses several proposals for collective licensing to enable development of generative AI systems while providing some compensation to copyright owners. It assesses the complex normative, economic, and practical problems that must be addressed if such a regime is to become feasible. It discusses the implications of a licensing mandate not only for the large firms whose models are widely used today, but also for start-ups, research centers, and higher education developers of generative AI systems, as well as the general public.”

5. 5

   Legal frictions for data openness: Reflections from a case-study on re-use of the open web for AI training

   HAL (Le Centre pour la Communication Scientifique Directe) (Ramya)Published Mar 27, 2025Checked Oct 9, 2026

   “While techno-legal openness is necessary, this report argues that the political economy of data re-use also necessitates legal strategies that impose certain limits on data extractivism by well-resourced actors like Big Tech on the one hand, and enable community data sovereignty on the other hand. It contains a repository of 41 ongoing legal controversies relating to copyright and data protection related to training foundation AI models, together with a detailed analysis of how these legal controversies either impact or advance three-dimensional data openness of training datasets. It also contains a critical analysis of existing open licenses, permissive licenses, as well as certain alternative licensing frameworks for training datasets. While these licensing frameworks impose more obligations on re-users and necessitate more collective thinking on interoperability, these licensing frameworks together with other legal and institutional changes are nonetheless necessary for the creation of healthy digital and data commons, to realise the original promise of the open web as open for all.”