The EU’s Link Tax Comes for AI Chatbots: IP and “Press Publisher’s Right”

by Stephan Kinsella on November 6, 2026

Matthew Broersma, “EU Court Adviser Says AI Summaries May Violate Copyright,” Silicon UK (Nov. 4, 2026); Eileen McDermott, “CJEU Advocate General Outlines Copyright Limits for AI News Summaries,” IPWatchdog (Nov. 4, 2026).

In a nonbinding opinion this week in a Hungarian publisher’s suit against Google over its Gemini chatbot, CJEU Advocate General Maciej Szpunar advised that when a chatbot retrieves and reuses a news article to answer users, “the press publisher’s right may be infringed in certain circumstances”—for instance, if it gets around a paywall, or if its summaries reuse “the form of expression of that publication in a recognisable manner,” so that users have no need to read the original. To be fair, he also said that facts presented in a substantially different form are not reproductions, and that training AI models on copyrighted works can fall under the EU’s text-and-data-mining exception. But note what is happening: the “press publisher’s right,” created by Article 15 of the EU’s 2019 Copyright Directive (the so-called “link tax”), is a recent invention designed to let newspapers charge Google for showing snippets of their stories, and now it is being stretched to cover AI answers that tell people what the news says. Under this theory, a summary becomes suspect precisely when it is good enough to be useful—that is, when it competes with the original. The complaint, in other words, is not copying but competition, which makes this a European cousin of the old “hot news” doctrine. This is copyright doing what it always does: giving one party a state-granted right to control how others use their own computers, servers, and words. The solution is not to fine-tune the EU’s rules on paywalls and “recognisable” expression, but to abolish copyright and its ever-multiplying offspring, the press publisher’s right included.