Overview
- Advocate General Maciej Szpunar delivered his opinion on Tuesday, Nov 3, 2026, concluding that AI chatbot answers can, in specific narrow cases, infringe press publishers’ copyright.
- Szpunar set two central tests for infringement: whether the chatbot accessed an article unlawfully, for example by bypassing a paywall, and whether the answer reproduces the publication’s distinctive expression so fully that users no longer need to consult the original.
- The opinion stresses that copyright protects an article’s specific expression not the underlying facts, so mere reporting of facts by a chatbot will not normally amount to infringement.
- Questions about whether Like Company’s articles were used to train Google’s Gemini were treated as hypothetical because no evidence was shown, and Szpunar noted that the EU Text‑and‑Data‑Mining (TDM) exception can permit some training uses.
- The EuGH opinion is non‑binding but influential; EuGH judges must now rule and the Budapest court has paused the Hungarian lawsuit pending that binding judgment, which could reshape how AI providers access paywalled content and manage liability.