The image we have of censorship is an old one: an official who reads, strikes out, bans. An 'unauthorised' stamp on a script, a work removed from a wall, a book that never goes to print. The censor is visible, the ban explicit and therefore contestable: there is something to point to, someone to accuse, an act to denounce.
The annual Freemuse report on artistic freedom, published in April 2026, documents a form of censorship that functions differently, harder to fight precisely because it barely resembles the earlier one. It is a qualitative piece of human rights monitoring that gathers cases from 2025 and cautions that it shows trends, not a complete census of the world. It offers no global figure and does not prove a statistical increase. What it offers, and it is enough to think with, is a pattern shown through cases and repeated in very different countries.
The pattern is as follows: governments no longer ban art so much as reclassify the artist. They do not say 'this song is banned'; they say 'the person who sings it is a foreign agent'. They do not say 'this film cannot be shown'; they say 'its director is a terrorist'. They do not strike out the work: they alter the legal status of its creator, and from that point on, there is no need to censor anything, because the ordinary system (criminal, tax, administrative) takes care of the rest by itself.
The report documents three tools, none of which are presented as censorship. Foreign agent laws, which designate artists and cultural organisations receiving external funding as security threats. Anti-terrorism legislation, applied to creators in countries as diverse as Turkey, Myanmar, or the United Kingdom. And morality, obscenity, or blasphemy regulations, which, as Freemuse describes them, rest on deliberately vague concepts: obscene content, degrading lyrics, promotion of crime. Vagueness is not a defect of these regulations: it is their function. An imprecise concept can be applied, in the report's words, selectively and arbitrarily against whoever is to be silenced.
Consider the elegance of the mechanism, if one may use such a term for this. Banning a work is an act of censorship; censorship has a bad press, mobilises defenders, is taken to international courts, and tarnishes the image of those who exercise it. However, prosecuting a citizen for violating anti-terrorism legislation is not censorship: it is the application of the law. The government does not appear as a censor; it appears as a guarantor of order. The burden of proof is reversed: it no longer has to justify why it bans a work; it is the artist who must demonstrate that they are not what the category claims they are. And while they demonstrate this, if they are able, they have already lost the work, the venue that was to host it, the funding, and at times, their liberty.
The British case detailed in the report is revealing precisely because it does not end in a conviction. There were terrorism investigations into musicians, there were charges, there were concert cancellations, yet the report clarifies that there was no criminal conviction. And yet, the effect occurred all the same, because it did not depend on the conviction. It depended on the reclassification: once an artist is tainted by suspicion of terrorism, venues cease to book them, festivals withdraw them, and sponsors distance themselves. Nothing had to be banned. It was enough to change the category under which they are viewed and allow everyone else to act accordingly. The report itself articulates this: once in motion, the mechanism narrows not only what an artist feels capable of saying, but what venues feel capable of hosting.
What is documented are the visible acts: the arrest, the charge, the cancellation, the name on a list. Yet the most effective part of this type of censorship is that which leaves no trace. The project that is not presented because the problem is anticipated. The programme that is modified before anyone objects to it. The work that is not created. When the tool is a vague legal category that could fall on anyone, it is not necessary to apply it often for everyone to apply the criterion to themselves. The most accomplished censorship is the one that gets the censored to censor themselves, and no report can record that, because there is no act to register.
It is not that classic prohibition has disappeared; it continues to exist, brutal, in many places. But it has become clear that the sophisticated form of cultural control no longer needs to ban. Controlling the categories is enough: who is an artist and who is a foreign agent, what is a song and what is the promotion of a crime, where the work ends and terrorism begins. Whoever controls those definitions need not censor a single work. They decide, before and above any specific work, what shall count as lawful culture and what shall count as something else, open to prosecution.
The question of artistic freedom, therefore, is no longer merely whether one may say such a thing. It is who possesses the power to decide that what you say is not, for legal purposes, art, but a crime. That decision is taken before you open your mouth and is far more difficult to appeal than a ban, because there is no ban to appeal. There is only a category, waiting.
About the open conversation
This text discusses the report 'The State of Artistic Freedom 2026' by Freemuse, a qualitative human rights monitoring work that documents cases from 2025 and which, as the organisation itself warns, shows trends rather than a global census. I am interested in a shift that the report reveals: contemporary censorship operates less and less by banning works and more by reclassifying artists through vague legal categories (foreign agent, extremist, obscene content), such that it is the ordinary penal or administrative system, and not a visible censor, that produces the silencing. I add, following a critique of the report itself, that the most effective part of this mechanism is that which leaves no documentary trace: anticipated self-censorship. If anyone wishes to intervene from the perspective of law, cultural management, or human rights, this notebook remains open.
Sources
Freemuse, 'The State of Artistic Freedom 2026: Courage Is Contagious: Art Under Fire', published on 17 April 2026 (freemuse.org). It documents the use of foreign agent laws, counter-terrorism legislation, and norms of morality, obscenity, or blasphemy against artists and cultural organisations, with cases in Russia, Turkey, Myanmar, the United Kingdom, Nigeria, Ethiopia, Pakistan, and other jurisdictions. Quote from the report: governments resort to 'foreign agent laws, institutional defunding and counter-terrorism charges' directed against artists and cultural workers. The report is qualitative and advocacy-based, not a statistical census; Freemuse warns that it should be understood as 'evidence of trends… rather than a comprehensive global record'. It does not provide a consolidated global total of cases nor a homogeneous year-on-year comparison.
In the cited British case (terrorism investigations against musicians), the report notes that there was no criminal conviction in the period described, but rather charges, cancellations, and institutional effects.
The observation that self-censorship and projects never presented leave less documentary trace than visible acts draws on a published qualification of the approach taken by this kind of report (Mimeta).