AI Cannot Build Its News Services at Journalism’s Expense – EPC Welcomes Recognition of Publishers’ Rights
The case addresses a fundamental question for the future of primary source journalism: when AI services use publishers’ reporting to answer users’ questions, what protection do publishers have over the work that makes those answers possible?
The European Publishers Council (EPC) today welcomed important elements of Advocate General Szpunar’s opinion in Like Company v Google Ireland, while questioning whether part of his proposed interpretation would impact publishers’ ability to control and license the use of their journalism by AI services. Advocate General Maciej Szpunar delivered his opinion in Like Company v Google Ireland, Case C-250/25 at the European Court of Justice today, 3rd November 2026.
The case addresses a fundamental question for the future of primary source journalism: when AI services use publishers’ reporting to answer users’ questions, what protection do publishers have over the work that makes those answers possible?
Every day, publishers invest in journalists, verification and editorial expertise to produce reliable information. AI services increasingly draw on that work to deliver answers directly to users, often removing the reason for readers to visit the original reporting. It is essential that the legal framework ensures that these substitutive products do not undermine the financial sustainability of the journalism on which they depend.
Commenting on the opinion, EPC Executive Director Angela Mills Wade said:
“The Advocate General recognises several important principles: making journalism freely accessible does not mean consenting to its use by AI chatbots, and simply rewording an article does not necessarily remove the need for a publisher’s authorisation. His reasoning also supports the recognition of clear rights reservations in website terms, including those expressed in natural language, which modern AI technology is capable of understanding.”
The EPC welcomed his view that AI services must not circumvent effective technical measures, such as paywalls or bot-blocking controls, to access journalism for grounding and publishers should be able to enforce those restrictions without sacrificing their visibility in search.
Angela added: “The Opinion also exposes a fundamental problem for rights holders: how can publishers enforce their rights when they lack the information needed to establish whether and how their journalism has been used? The absence of evidence of use is not necessarily evidence that no use has occurred. Effective copyright protection requires meaningful transparency, traceability and the ability to verify how protected content is exploited by AI services and this is exactly what EPC has been calling for under the AI Act and its Code of Practice.”
EPC supports a market for voluntary, negotiated licensing that enables AI innovation while preserving publishers’ control over how their content is accessed and used, and to determine its value. Effective protection must address the commercial reality of services that use journalism to attract and retain audiences, and give publishers a meaningful basis on which to negotiate.
The stakes extend beyond individual articles or technologies. If AI services draw value from original reporting while weakening the businesses that fund it, the public ultimately loses access to reliable, independent news. Europe’s commitment to journalism must be reflected in rights that work in practice.
The opinion was delivered on 3 November 2026 in Case C-250/25, Like Company v Google Ireland. It is advisory and does not bind the Court of Justice.
The Advocate General draws an important distinction between the facts reported in journalism, which are not themselves protected, and the specific expression of a press publication, which is. He considers that a chatbot's retrieval and reuse of press content may infringe publishers' rights, particularly where it circumvents effective technical measures restricting access or scraping, or where its answers reproduce the publication's expression recognisably and so extensively that users no longer need to consult the original reporting.
He further considers that the making-available right is, as a rule, infringed where that protected expression is systematically recognisable in answers to related queries. Importantly, the chatbot provider may itself be liable where such uses are inherent to its business model or it has failed to introduce credible and effective measures to prevent them.
These conclusions recognise that AI-generated answers cannot necessarily escape publishers' rights simply by reformulating their journalism, and that responsibility for unlawful exploitation may rest with the AI service itself.
