"AI Reuses Articles' Distinctive Wording... Rights Infringed if Readers No Longer Need to Read Original"
EU Advocate General: "May Constitute Copyright Infringement"
Legal Battles Underway Around the World, Including in the U.S. and India
The Court of Justice of the European Union (CJEU) has issued an opinion saying that a news publisher’s copyright may be infringed if generative artificial intelligence (AI) reuses wording from an article to provide a summary, making it unnecessary for readers to read the original article separately.
Maciej Szpunar, an advocate general at the CJEU, issued the opinion on November 3, 2026, local time, in a neighboring-rights infringement case (C-250/25) brought by Hungarian media company Like Company against Google Ireland. It is the court’s first case concerning generative AI and EU copyright law. If the court adopts the opinion, it could serve as a standard for courts in EU member states deciding similar disputes.
Like Company sued Google in a Hungarian court, alleging that Google’s Gemini AI had used its articles without permission for training and reproduced them in its responses. In April 2025, the Hungarian court asked the CJEU to interpret the law.
Reusing Original Text in Summaries May Be a Problem
The key issue is whether AI’s use of news articles infringes publishers’ neighboring rights. In 2019, the EU granted publishers of press publications, under its Copyright Directive, the right to control the reproduction and online availability of articles, separate from journalists’ copyrights.
Szpunar proposed the method of access to an article and the extent to which the response reuses its original text as criteria for determining infringement. He said that circumventing a paywall for a paid article or technical measures designed to prevent automated scraping could be considered unlawful access.
He also said that summarizing facts contained in an article alone does not constitute infringement. However, infringement may occur if the original article’s sentences or distinctive expressions are reused to a recognizable extent, making readers feel no need to read the original article separately. He added that repeatedly providing similar responses could also raise issues of rights infringement.
In particular, he said that an AI company could be held liable if reproducing wording from original texts is part of its business model or if it has failed to put measures in place to prevent such reproduction.
Szpunar said that, in this case, the question of whether infringement occurred should exclude the claim that Like Company’s articles were used for training, as there was no evidence that they had actually been used. Considering how Gemini accessed the articles and the extent to which its responses reused their wording, he offered the view that it would be difficult to find Google liable for infringement. An advocate general’s opinion is not legally binding.
U.S. Department of Justice: "AI Training May Qualify as Fair Use"
Debate is also underway in the United States over the legality of AI training and response generation. In 2023, The New York Times sued OpenAI and Microsoft, alleging that they had used its articles without permission to train AI and that ChatGPT and other services reproduced article content, becoming substitutes for news services.
In September 2026, the U.S. Department of Justice submitted an opinion in the lawsuit saying that the copyright law principle of "fair use" could apply to AI training. Its position was that learning language patterns by training on articles serves a different purpose from reproducing and providing the original text, but that the reproduction of protected expression in responses should be assessed separately. Recently, Hollywood and music industry groups opposed unauthorized AI training and submitted a brief in support of The New York Times.
In July 2026, the Delhi High Court in India rejected an interim injunction application in a copyright infringement lawsuit brought by news agency ANI against OpenAI. It provisionally ruled that storing articles for training fell under the "fair dealing" exception, and found that it had not been sufficiently demonstrated that ChatGPT had reproduced protected expression from the articles in a substantially similar form. ANI appealed, and hearings are ongoing.
Hot Picks Today
"Two Spend 500,000 Won in a Flash, Yet Wait Two Hours?" ... 'Dreamland' Shunned by Kids and Foreign Tourists
- The Era of Getting 200 Million Won After 30 Years Is Over... A Japanese Bank's "Perpetual Department Head" Who Just Clocked In Cries Out
- Spain Hit by Its Worst Tragedy... Lee Kangin, Moreno and Real Madrid Join Soccer World in Mourning
- They Said, “Make It Your Home for Life” - Then Suddenly Kicked 200 Passengers Off... What Happened to the Around-the-World Cruise?
- "130 Rice Bowls Sold Out in a Day"... 'Unfiltered Face' Adds 8 Million Won to Night Market Sales: China's Bungeunwoo
Jo Hanju, reporter for the Legal Times
※This article is based on content supplied by Law Times.
© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.