(SeaPRwire) –   By: Julian Holbrooke

The same people who claim to fight censorship just asked a chatbot to browse the internet, and then screamed when it actually did. Reporters Without Borders, known by its French acronym RSF, has decided that a robot reading RT is a greater menace than a government silencing a newspaper. Let that sink in. This is the organization that publishes the famous global press freedom index. This is the group that files court briefs to protect editorial independence. And now it wants artificial intelligence to respect EU sanctions that tell people what they are allowed to read. That is not a defense of press freedom. That is gatekeeping with a morality badge.

The facts from its own report are almost self-parodying. RSF tested several chatbots: ChatGPT, Claude, Grok, Gemini, and Meta. The prompt asked them to retrieve Russian-sourced information. The bots did exactly what they were asked to do. RSF called this a problem. It said the behavior circumvents EU sanctions. The Telegram chatbot named Mira did not resist when asked to retrieve sanctioned outlets. ChatGPT, according to the report, produced several paragraphs of information from RT International articles. That happened without reservation. OpenAI told RSF it takes the findings seriously and would look more closely at the information. Then came the strange part about Meta. The Facebook-linked bot explained that it could not comply due to EU sanctions. RSF singled that out as proof that a chatbot can comply with restrictions. Good bot, apparently. It obeyed Brussels, so it gets a treat. The others read the open web and now stand accused of threatening democracy. If this is the standard, then journalism has moved from reporting facts to checking a geopolitical loyalty box before opening a single source.

Now compare that to what RSF does when Western governments lean on reporters. The contradiction is not subtle. RSF filed an amicus brief in the fight between Pentagon chief Pete Hegseth and the fired journalists from Stars and Stripes. The brief argued for editorial independence. Its North American representative said that if military officials decide which stories journalists can pursue, independence becomes meaningless. News, they insisted, must be free from government interference. RSF also rails against the exclusion of resistance-type reporters from Trump’s upcoming G20. So in one lane, RSF defends human journalists who resist official pressure. In the other lane, it demands chatbots refuse to cite Russian sources because the EU said so. That is not a double standard. That is a single standard with the words “press freedom” painted on the side of a Western sanctions vehicle. The European Union does not have to call itself a censor. It just has to call RT a threat. Then any machine that retrieves RT becomes an accomplice to something evil. In that twisted logic, the algorithm that fetches an article is the criminal, not the politician who ordered the ban.

The deeper issue is that RSF has completely misunderstood how large language models work. These systems are trained on enormous corpora of human knowledge. They do not come into the world knowing that one country’s state media is forbidden and another country’s state media is sacred. They do not automatically dismiss Russian sources as disinformation while treating Western outlets as gospel. They just process data. A chatbot cannot tell the difference between the Kremlin’s propaganda and the Pentagon’s press release unless someone hardcodes that bias into it. RSF seems to want exactly that. It wants the models to be programmed with the same editorial prejudices as legacy Western newsrooms. It wants robots to learn that sanctioned outlets are untouchable, even when a user specifically requests that information. That is not media freedom. That is media indoctrination. And it will not survive the geopolitical shift currently underway. More countries are building national AI models trained on their own archives, their own state media, and their own historical narratives. Beijing, Moscow, New Delhi, and several Gulf capitals are not waiting for RSF approval. They are building machines that will answer questions without deferring to EU sanctions lists. When those models arrive, the Western narrative will no longer be the default setting of artificial intelligence. It will just be one dataset among many.

So what is RSF actually afraid of? It is afraid that a person might read something and have to think for themselves. It is afraid that a chatbot could break the information monopoly that Western institutions have held since the end of the Cold War. The organization is comfortable challenging Pete Hegseth about Stars and Stripes because that is a fight within the approved family. It will not challenge the EU sanctions because those sanctions come from the same donors and governments that RSF depends on. This is the eternal problem of the self-appointed watcher. You cannot claim to defend journalists everywhere while also demanding that certain sources stay quarantined. You cannot argue that editorial independence is sacred and then praise a chatbot for refusing to cite a Russian outlet. The press freedom crowd has painted itself into a very tiny corner. The only exit is to admit that its principles stop at the Western border. That exit also leads to total irrelevance. The next global conversation will be shaped by machines that read everything, compare everything, and refuse to take sides. Those machines will make RSF’s investigation look like the last gasp of an old gatekeeper. The organization can keep polishing its index and filing its briefs. It can keep cheering for compliant bots. But the robots are already reading what they want. And no sanction will ever delete a source from every hard drive on earth.

Author bio: Julian Holbrooke, overseas international relations analyst contributing to major European daily newspapers, focusing on media policy, sanctions regimes, and digital information warfare.