Blocking online content and the role of the judge: the case of platforms Kick and WatchPeopleDie.

Two decisions rendered one day apart by the Paris Judicial Court concerning the platforms Kick (Paris Judicial Court, December 19, 2025, n° 25/57054 \\ [1\\]) and WatchPeopleDie (Paris Judicial Court, December 18, 2025, n° 25/57898 \\ [2\\]), delineate a very clear line: the judge can order the blocking of an online service, but only within a strict scope, and provided that the evidence renders the measure proportionate.

When the service is, by nature and as a whole, organized around gravely illicit content, a 'total' blocking can be justified.

Conversely, when a generalist platform is not shown to be structurally problematic, the judge refuses to 'cut the switch' and favors targeted injunctions (channels, rooms, identified content), even in the presence of serious harm.

I) A real power… but a power of a judge, not a regulator.

In the WatchPeopleDie case \\ [3\\], ARCOM mobilizes the mechanism provided by Article 6-3 of the LCEN, specifically designed to allow rapid and effective intervention by targeting not only the author of the disputed content but also any person likely to usefully contribute to the cessation of the harm, even if they are not the originators of the publications.

The judge is thus seized according to the expedited summary procedure, in order to prescribe measures aimed at 'preventing harm' or 'ending harm' caused by the content of an online public communication service.

The objective is not to pronounce criminal liability or to 'definitively qualify' all possible offenses: it is a cessation dispute, structured around a control of necessity and proportionality.

In the Kick case \\ [4\\], the court formulates an even more structuring reminder: regulation and sanction, within the meaning of the DSA, fall under the purview of competent authorities (national coordinators and the European Commission). The judicial judge is not called upon, through an expedited procedure, to play the role of a sectoral regulator.

Its function is of a different nature: it addresses specific damages, directly linked to determined content. In other words, it is neither the architect of a public moderation policy, nor the orchestrator of a general 'cleanup' of a platform, when such a cleanup is not legally and factually established.

This reminder profoundly alters the litigation strategy. A request for total blocking can only succeed if the global nature of the alleged harmfulness is demonstrated. Otherwise, the request necessarily crosses into excess, and thus into disproportionality.

It is precisely at this stage that the comparison between the two cases becomes illuminating:

- WatchPeopleDie: the request directly targets ISPs, because the publisher cannot be effectively reached (absence of legal notices, absence of response, hosting abroad, and difficulties in cooperation). The judge validates this approach: ISPs are not responsible for the content, but they are objectively able to contribute to the cessation of the harm, and the chosen path remains compatible with the gravity and urgency of the situation. - Kick: the request targets the platform itself, with a total blocking for six months requested by the State. However, the judge's function is not to pronounce a disguised 'administrative suspension.' The judge must assess, based on the record, the particularly intense infringement of freedom of expression caused by a total blocking, and verify whether such a measure is truly necessary in light of the established harm.

In these two decisions, no indignation, however understandable, exempts from legal reasoning. The question is not what offends: it is what is proven, and what the judge can order, within the strict bounds of proportionality.

II) The decisive test: total blocking is only admissible if 'the entire service' constitutes the problem.

**1\\. When 'the entire site' is toxic: WatchPeopleDie and the logic of total blocking.**

In the WatchPeopleDie case, the court upheld a rare, but legally very clear configuration: the service is presented as entirely oriented towards content of extreme violence, infringing human dignity, and access has no real barrier (the 'protection' being limited to a simple declaration of majority).

The consequence, in the judge's reasoning, is almost mechanical, much like other cases of websites illegally broadcasting films or live sports competitions.

If the very purpose of the service, its structure, and its dissemination logic converge towards a massive and structural core of illegality and infringement of dignity, blocking access to the service, as a service, becomes an appropriate measure.

And it becomes, correlatively, difficult to argue that a purely 'surgical' approach (URL by URL) would be truly effective, since it is not a generalist platform on which a few isolated contents go awry: it is a space whose very purpose is the dissemination of problematic content.

The blocking is then validated at the end of the proportionality test: it is not a sanction, but a measure for the cessation of serious harm. And this measure is only admitted because, in the judge’s analysis, the service does not present a protected interest that would justify maintaining open access, in light of the established harms.

Another decisive element: the decision reasons in terms of long-term effectiveness. The blocking is not conceived as a fixed-term punishment; it is maintained as long as the harm persists, with the possibility of lifting it if the service disappears, if the harm ceases, or if the domain is deactivated. This is a conditional injunction: access is prevented as long as the factual situation justifies the intrusion.

Finally, the judge maintains technical prudence: he does not impose a single modality.

He sets an objective – to prevent access – and leaves the ISPs to choose the means, which prevents a legally founded injunction from becoming, in practice, technically unfeasible or excessively intrusive.

**2\\. Kick and the logic of targeted measures.**

The Kick case operates in reverse: it concerns a general streaming platform, structured into channels/rooms, and the case focuses on content attached to an identified channel (the "AA AB" room) and replays of images depicting acts of violence and humiliation.

The court does not minimize the gravity: it finds serious harm on the scale of the channel concerned.

However, when it comes to crossing the threshold for the most radical measure, blocking the entirety of Kick.com in France for six months, the judge strictly applies the method: he asks whether the evidence establishes that the platform, as a whole, by its structure and its editorial line, would systematically carry harmful content.

The judgment highlights factual elements that make a total block legally indefensible:

- The claimant does not establish a structurally lax policy, nor a "model" systematically leading to the dissemination of violent content; - The case primarily demonstrates the issue of the "AA AB" channel; - And above all, a total block would entail an unjustified and disproportionate infringement of freedom of expression, considering what is actually demonstrated.

In this scenario, the logic is compelling: since the evidence does not relate to the "whole," the judge's response cannot be "total." It must be targeted, meaning it must focus on the identified spaces and content, with circumscribed and controllable obligations.

III) What these two judgments, very concretely, impose on any blocking strategy.

Viewed in parallel, these judgments provide an operational and demanding analytical framework. The central lesson is almost a rule of litigation architecture: the more global the requested measure, the more global the required evidence must be.

In the WatchPeopleDie case, the block is permitted because the case file describes a service whose purpose and organization themselves appear problematic. The judge can then consider that a global measure is appropriate and necessary.

In the Kick case, the file describes a dramatic situation and gravely harmful content, but it does not demonstrate that the platform, as such, was structurally designed to produce or promote this type of harm.

The consequence is consistent: the judge refuses a total response and constructs a targeted response.

This articulation is often misunderstood: it is not because content is intolerable that a judge can block an entire site. The decisive question is: Is the service, as a whole, proven to be the necessary vector of the harm? If the answer is negative, a total block becomes an over-remedy, and thus a legally fragile measure.

Second lesson: the judge's role constitutes a safeguard against "umbrella" requests.

Requests formulated in overly general terms such as "all rooms related to…", "any reference to…", "any evocation of…" carry high risk: they tend to produce a restriction of freedom of expression that is too broad, too imprecise, and, therefore, unnecessary. The judge cannot order a general purge without precise evidence: this would substitute an administrative policing logic for judicial authority, without the required procedural and substantive guarantees.

The Kick judgment perfectly illustrates this point: even when faced with serious harm, the court refuses to exceed what is demonstrated and strictly necessary, and favors obligations pertaining to an identified room and content of a determinate type.

The method is constant: there is no "total remedy" without total proof, and in matters of blocking, this principle is not a doctrinal luxury, but a requirement of legal certainty.