When a Social Media Post Becomes a Copyright Work: What C-598/24 Means for Online Publishing, AI and Content Governance

On 3 September 2026 the Court of Justice of the European Union gave judgment in Case C-598/24, Gândul Media Network (ECLI:EU:C:2026:682), on a deceptively simple question: can a short text posted on social media be a copyright-protected work?

The Court's answer was yes, provided the text is the expression of an intellectual creation reflecting the personality of its author. Whether a particular text meets that test remains for the national court to decide.

That conclusion matters well beyond the facts of the case. It applies settled EU copyright law: a text's length, the fact that it was posted online and whether it belongs to a recognised literary genre do not decide whether it is protected, unless those factors left the author no room for creative freedom. What decides is originality. For organisations that publish, repurpose, summarise, or use AI systems to process digital content, the judgment is a reminder that informal online material can still carry legal rights.

The dispute

On 8 September 2021 a Romanian teacher published on her Facebook page a 22-line text entitled "Short guide for parents at the start of the school year", saying, in essence, that she did not wish to receive gifts from her pupils' parents. Six days later a journalist published an article on the website of the online daily newspaper Gândul that reproduced the entire text without her prior consent. According to the order for reference, her name and a hyperlink to the Facebook page were added subsequently (paragraphs 9 and 10).

She sued the publisher and the journalist for copyright infringement. The Regional Court in Bucharest dismissed the claim in 2022 on the ground that the text was not eligible for copyright protection, and the Court of Appeal upheld that ruling in 2023. The High Court of Cassation and Justice then asked the CJEU whether such a post can be a "work" under Article 2(a) of Directive 2001/29 and, if so, how far the exception for reporting current events reaches (paragraphs 11 to 16).

A social-media post can be a copyright work

The Court restated the two cumulative conditions for a "work": the subject matter must be original, in the sense of being the author's own intellectual creation, and it must be the expression of that creation. Originality exists where the subject matter reflects the personality of its author, as an expression of free and creative choices (paragraphs 19 and 20).

The Court held that a text's length, the fact that it was posted online, and whether it belongs to a recognised literary genre are irrelevant to that classification, unless those factors result from rules, technical considerations or constraints that left the author no room for creative freedom (paragraph 25). Originality can arise from "the choice, sequence and combination of the words" by which the author expressed her creativity (paragraph 26). On the information before it, the Court observed that the teacher appeared to have made free and creative choices, subject to the national court's assessment (paragraph 27).

That is the central legal point. A social-media post is not outside copyright merely because it is short, was posted online, or belongs to no recognised literary genre. What matters is whether it is the author's own intellectual creation.

Public does not mean unprotected

There is an operational lesson here. Digital availability is often confused with legal availability. Content may be publicly accessible, indexed by search engines, widely shared, easy to copy, embedded in a public platform and technically downloadable, and still be protected by copyright.

That distinction grows in importance as organisations automate content collection, summarisation, media monitoring, knowledge management and AI-assisted research. The governance question is not only whether a system can access content. It is also on what legal basis the organisation may reproduce, transform, retain, redistribute or commercially use it. C-598/24 does not create that distinction, but it illustrates it clearly.

Reporting current events is not an unlimited reproduction right

The second part of the judgment concerned the exception in Article 5(3)(c) of the Directive, which allows Member States to permit the "use of works or other subject matter in connection with the reporting of current events, to the extent justified by the informatory purpose and as long as the source, including the author's name, is indicated, unless this turns out to be impossible".

The Court said it cannot be ruled out that a newspaper's use of the teacher's post was reporting a current event, leaving that determination to the national court (paragraphs 37 and 39). The functioning of the school system may be of public interest, and "reporting" requires neither a detailed analysis of the event nor an invitation to the public to react (paragraph 38).

The exception is, however, bounded by proportionality. The Directive sets no formal limit on the length of the work used, but the use must not go beyond what the informatory purpose requires (paragraph 49). Applying the three-step test in Article 5(5), except for a text so short that using extracts is impossible, reproducing a work in full may conflict with its normal exploitation and unreasonably prejudice the author's legitimate interests, because the reproduction replaces the original communication and spares readers from consulting it (paragraph 53). The Court did not decide whether the newspaper's reproduction was lawful; that remains for the Romanian courts.

So the compliance question is not only "did we credit the author?" It is also "did we use more of the work than the reporting purpose needed?" Attribution and permission are not interchangeable.

National law cannot exclude commercial use from the exception

Romanian law allowed short extracts to be reproduced only where no direct or indirect financial or commercial benefit was derived. The Court held that a Member State may restrict the exception to short extracts, provided the restriction is proportionate and keeps the exception effective (paragraph 54). But a general prohibition on deriving a commercial or financial benefit has no basis in the wording of Article 5(3)(c), whose informatory-purpose condition concerns only the extent of the use (paragraph 55). Media outlets, the Court noted, inform the public and scrutinise public figures while also pursuing the economic activity they need to operate, and such a prohibition would undermine the exception and upset the balance between intellectual property (Article 17 of the Charter) and freedom of expression and of the press (Article 11) (paragraph 56).

The judgment is therefore neither a simple victory for authors nor for publishers. It reinforces copyright protection while preserving real space for freedom of information and press activity.

Why this matters for AI governance

The Court was not deciding an AI case. C-598/24 says nothing directly about whether AI-generated text attracts copyright, whether model training infringes, or how authorship is allocated where generative AI is involved.

Its reasoning is still relevant, because it restates the originality test in strongly human terms: the author's free and creative choices, the expression of personality, the choice and arrangement of words. That raises a governance question for organisations that use generative AI: where, in an AI-assisted production process, are the legally relevant human creative choices actually made? The answer may differ considerably between a person writing independently, a person using AI for grammatical editing, a person directing a model through detailed iterative prompts, a person choosing among generated alternatives, and a system that generates and publishes autonomously. Those scenarios should not be treated as equivalent.

In practice this becomes a question of provenance. If rights depend partly on human creative contribution, organisations may need to record who initiated the work, whether and how AI was used, what material human edits occurred, who approved the final version, what source material was incorporated, and what permissions or licences applied.

Content governance has to move upstream

Many organisations still treat copyright as a downstream legal review: content is created or collected, published, and legal is consulted when a dispute arises. That model becomes fragile once content operations are automated, because AI systems can retrieve, summarise, transform and republish material at a scale no individual review can follow.

The response is not more lawyers at the end of the process. It is rules earlier in it: which sources automated systems may access; whether full-text reproduction is permitted; when a summary should replace reproduction; when attribution is required; whether licensing restrictions apply; whether material can be reused commercially; whether human review is required before publication; and how provenance and permissions are recorded. Copyright obligations need to become operational controls, not only legal advice.

The broader compliance lesson

A compliance conclusion rarely follows from a single rule. Here the legal position turns on the nature of the material, whether it was a protected work, how much of it was reproduced, the purpose of the reproduction, whether it reported a current event, whether the source was identified, and whether the national implementation stayed within what EU law permits. That is why compliance does not reduce to a static checklist. The sequence is closer to context, then applicability, then obligation, then control, then evidence: only once the facts are established can the applicable requirements be determined.

For a product-design view of the same originality test, see When Product Design Becomes Copyright: What the Birkenstock-Lidl Case Means for IP Governance.

What organisations should take from C-598/24

The immediate lesson is simple: do not assume that informal digital content falls outside copyright. A short social-media post can be a protected work if it reflects its author's free and creative choices.

The strategic lesson is broader. As organisations use automated systems to create, collect and reuse content, they need governance that can answer three questions: what is this content, what rights attach to it, and what are we permitted to do with it? Those questions sit where intellectual property, AI governance, data governance and digital publishing meet. As C-598/24 shows, the legal consequences can begin with a 22-line Facebook post.

How Priventia approaches it

Priventia's governing principle is that every compliance conclusion should be explainable. Its method links the legal requirements that apply to an organisation to the controls that address them, the evidence behind those controls and the monitoring that keeps them current, with every step traceable to the provision it comes from. Intellectual property is one of Priventia's regulatory domains, with copyright as its first family, and its coverage will be built on the same discipline.

Source: CJEU (Second Chamber), Case C-598/24, CY v Gândul Media Network SRL and HO, judgment of 3 September 2026, ECLI:EU:C:2026:682 (Publications Office text). Paragraph references are to the judgment.