# AI

By [DYLIT Media Buzz](https://dylit.info/user/dylitmediabuzz)

[Global Finance](https://dylit.info/pr/global-finance/6a6849fd61bb0e07120aa486) > [AI & its Impacts](https://dylit.info/ch/ai-its-impacts/6a6849fd61bb0e07120aa4a2)

Copyright in the AI Era: Why Borders Still Matter When Borders Blur: Jurisdiction and Enforcement Artificial intelligence is inherently transnational. Training data is scraped from servers scattered across dozens of countries, models are trained on cloud infrastructure that can sit anywhere, and the resulting tools reach users worldwide within seconds of release. Copyright law, by contrast, remains stubbornly territorial: each country applies its own rules about what counts as infringement and where. The UK's landmark Getty Images v Stability AI case shows how disruptive this mismatch can be. Getty's claim centered on allegations that Stability AI infringed copyright by storing images on servers in the UK while training its Stable Diffusion model, and the court ultimately found that training conducted abroad does not guarantee immunity from UK liability, though it can materially change a company's legal exposure. Because evidence of where training actually happened is usually held only by the company being sued, courts on both sides of the Atlantic are struggling just to establish the basic facts before they can apply the law. Ongoing US litigation, including The New York Times v. OpenAI and Microsoft, has faced similar friction over what evidence must be produced and from where. The Monetization Squeeze This jurisdictional patchwork weakens the leverage creators have traditionally relied on. A publisher, photographer, or musician who negotiates licensing terms — or wins an injunction — in one country cannot easily stop the same underlying model from being trained, hosted, or served to users from a jurisdiction with looser rules. A "no" in one market can effectively be routed around elsewhere. Meanwhile, AI-generated summaries and answers increasingly sit between publishers and their audiences, reducing the traffic and subscription revenue that content creators depend on, even when no single act of infringement can be pinned to a specific court's jurisdiction. The result is a structural imbalance: creators bear the cost of production locally, while the economic benefit of their work can be captured by systems trained and monetized somewhere else entirely, often beyond the practical reach of the courts where the harm is felt. A Handful of Gatekeepers? This dynamic feeds a broader concern: that a small number of well-capitalized AI companies, often controlled or founded by a handful of extremely wealthy individuals, could come to dominate how information is generated, filtered, and distributed globally. Because only a few firms have the capital, compute, and legal resources to train frontier models and defend the resulting lawsuits, jurisdictional complexity arguably favors incumbents — smaller publishers and independent creators often cannot litigate across multiple countries simultaneously. Critics worry this could concentrate influence over news, scholarship, and creative culture in the hands of a few companies whose commercial incentives may not align with pluralism, accuracy, or fair compensation for the humans whose work trained their systems. Open Questions of Principle and Ethics Beneath the jurisdictional and economic issues lie deeper, unresolved questions that courts and legislators have only begun to address. Is training a model on copyrighted work inherently "transformative," or does it functionally compete with and substitute for the original? US courts have split in early rulings, and the US Copyright Office itself concluded that some AI training uses are unlikely to qualify as fair use, while others may. Do creators deserve compensation simply because their work had value to a system, even if no identifiable copy appears in any output? What obligations, if any, do AI developers owe to disclose what they trained on, or to let creators opt out in advance rather than after the fact? These are not settled matters of black-letter law; they are live disputes over principle, currently being argued case by case, country by country, with genuinely reasonable positions on multiple sides. Why Deeper Scrutiny Is Needed Because these questions touch fundamental issues of authorship, fair compensation, market power, and free expression, they deserve more than piecemeal resolution through scattered private lawsuits and inconsistent national rules. Judicial scrutiny matters because courts are where the facts of training, storage, and distribution actually get tested — as Getty Images v Stability AI showed, much of the law's application still turns on unresolved evidentiary questions. Parliamentary and congressional scrutiny matters because only legislatures can build coherent frameworks for licensing, transparency, and cross-border enforcement that litigation alone cannot deliver. Until governments coordinate more deliberately — through bodies like WIPO or through harmonized national reforms — creators, AI companies, and the public will continue to operate in a legal environment defined more by jurisdictional accident than by settled principle. Sources U.S. Copyright Office, Copyright and Artificial Intelligence, Part 3: Generative AI Training — https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-3-Generative-AI-Training-Report-Pre-Publication-Version.pdf U.S. Copyright Office, Copyright and Artificial Intelligence, Part 2: Copyrightability — https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-2-Copyrightability-Report.pdf U.S. Copyright Office, AI initiative hub — https://www.copyright.gov/ai/ Directive (EU) 2019/790 on Copyright in the Digital Single Market (Article 4, text and data mining exception) — https://eur-lex.europa.eu/eli/dir/2019/790/oj WIPO Conversation on Intellectual Property and Frontier Technologies (AI) — https://www.wipo.int/en/web/frontier-technologies/frontier_conversation OECD AI Principles — https://oecd.ai/en/ai-principles Harvard Magazine, "Is Copyright Law the Wrong Weapon Against AI?" — https://www.harvardmagazine.com/five-questions/harvard-copyright-law-artificial-intelligence YouTube Videos U.S. Copyright Office webinar, International Copyright Issues and Artificial Intelligence — https://www.youtube.com/watch?v=QaUzkerRSdM WIPO Director General Daren Tang on IP and AI, IPWatchdog Unleashed — https://www.youtube.com/watch?v=GzZg0GHFJBk WIPO panel, The Right to Research in International Copyright — https://www.youtube.com/watch?v=_nhsi_WrYvk IPWatchdog, Artificial Intelligence, Fair Use and the Future of Copyright Law — https://www.youtube.com/watch?v=joDQJp-KM4g