Imagine a future where the government quietly decides which news stories are safe for you to see, which sources deserve to be promoted in your feed, and which opinions are pushed to the margins as “false and misleading.” That is no longer a dystopian thought experiment. The British Government is actively building the machinery to do exactly that, and it is dressed up as a public-spirited defence against online misinformation. The blueprint includes statutory “quality signals” that would let the state influence how content is ranked, a government-linked database of “false and misleading narratives,” official “prebunking” campaigns woven into schools and public communications, and a Green Paper proposal that would force social media algorithms to give prominence to news sources the state deems “trustworthy.” Ofcom, the communications regulator, even has a crisis protocol ready for the next national emergency. The official justification for all of this rests on two bold claims: that misinformation has saturated the internet and is causing grave harm, and that platform algorithms are the engines driving it. But when you actually test those claims against primary sources and peer-reviewed research, the whole evidential foundation begins to crumble. Before we let anyone build a Ministry of Truth, we should ask for stronger receipts than these.
The supposed “full evidence base” that ministers said they were missing was produced by a Government Office for Science review that cost taxpayers almost half a million pounds. You might expect that kind of investment to produce a careful, rigorous, and transparent analysis. Instead, the authors openly relied on AI-generated summaries of 228 papers they had not actually read. They also drew on monthly questionnaires filled in by fact-checkers affiliated with the European Digital Media Observatory. That is not exactly the gold standard of scientific inquiry. To make matters worse, the review’s crucial claim that platform algorithms are “exclusively profit- and engagement-driven” rests on a single citation from an obscure Indonesian institution ranked 10,370th in the world. If a plumber charged you half a million pounds to fix a leak and then handed you a report generated by ChatGPT from articles he never opened, you would probably feel cheated. But this is supposed to be the evidence underpinning some of the most sweeping restrictions on online speech in modern British history. The review also ignored contradictory peer-reviewed research that undermines its central premises. On top of that, the people who have been pushing hardest for these measures are a small, tightly connected network of campaign groups, consultancies, self-appointed fact-checkers, and former broadcasters whose recommendations always seem to lead in the same direction: more censorship. The Centre for Countering Digital Hate, which has lobbied relentlessly for tougher regulation, was co-founded by Morgan McSweeney, who went on to become Sir Keir Starmer’s Chief of Staff. After the Southport tragedy, that same group helped chair an “emergency” roundtable with officials from the Department for Science, Innovation and Technology, the Home Office, Ofcom, and the Metropolitan Police. The Global Disinformation Index has been funded by the Foreign Office. When the government needed a review to confirm that the problem was real, members of this same lobby network conveniently served as expert reviewers. It is a closed loop, and the taxpayer funded it.
What do the actual numbers tell us about the scale of misinformation? Far less dramatic than the alarmists would have you believe. A 2024 study published in Science looked at 2.7 billion vaccine-related views on Facebook during the rollout and found that misinformation flagged by fact-checkers accounted for just 0.3% of those views. That is a miniscule fraction. Meanwhile, “vaccine-skeptical” content that contained no outright falsehoods had 46 times the aggregate impact on vaccination intentions. And the single most influential item was not some wild conspiracy theory from the dark corners of the internet; it was an accurate article from the Chicago Tribune, seen by more than 50 million people. Think about that. The story that scared people about the vaccine was true reporting, not fake news. Another 2024 synthesis in Nature found that exposure to misinformation is far smaller than commonly assumed, with so-called “fake news” accounting for roughly 0.15% of the average American’s daily media diet. That means 99.85% of the content people consume is not fake news. If the enormous regulatory apparatus being proposed is meant to solve a crisis of this scale, the medicine seems wildly disproportionate to the disease. We are being asked to accept sweeping censorship powers to fix a problem that, according to the best evidence, is relatively small and not clearly driving the harms described. That should make any reasonable person pause.
The claim that social media algorithms are radicalising us by rewarding extreme content also falls apart under scrutiny. Twitter’s own researchers studied millions of accounts and found no evidence that its algorithm systematically favoured far-left or far-right voices over moderate ones. A landmark Facebook experiment found that the algorithm actually served less untrustworthy content to users than a simple chronological feed would have. In other words, the opaque, engagement-hungry algorithm was doing a better job at keeping junk away than a raw stream of posts would have done. The moral panic around algorithms amplifying conspiracy theories is not supported by the evidence. And what about the idea that belief in conspiracy theories is exploding across the population? A study tracking 46 different conspiracist beliefs in national surveys since 1966 found no evidence that belief in conspiracy theories is rising overall. In fact, significant declines outnumbered increases. That is the opposite of what you would expect if the internet were turning everyone into a tinfoil-hat radical. This matters because the Government is proposing to force platforms to boost “trustworthy” sources and demote or suppress material that falls outside the state-approved narrative. If the entire justification is built on a false premise about how algorithms work and how much misinformation people actually consume, then the entire intervention is fundamentally unjustified. Maybe the reason we keep hearing about this crisis is not because it exists, but because a well-connected campaign industry needs us to believe it exists.
When the Government starts talking about promoting “trustworthy” news, you have to ask: trustworthy according to whom? The state’s own favourite example of a reliable, impartial broadcaster is the BBC, but the public is not exactly convinced. Ofcom’s Media Nations 2026 report found that only 33% of adults associate the BBC with impartiality, and only 45% with trustworthiness. That means nearly two-thirds of the public no longer see the BBC as neutral. And recent events do not inspire confidence. Panorama was found to have doctored a Donald Trump speech, and Ofcom itself issued a critical ruling about a Gaza documentary. It is hardly a surprise that public trust in the state’s chosen arbiter of truth is so low. There is something deeply uncomfortable about the idea that the same government that brought us these institutions should now be handed the power to decide which news outlets are promoted and which are buried. The state has an interest in maintaining its own legitimacy, especially during crises. Granting it the power to define what counts as “misinformation” is like letting a referee mark their own homework. The history of censorship is full of examples where those in power called inconvenient truth “misinformation” and dissent “harmful content.” The British Government’s current plans, with all their talk of quality signals and crisis protocols, echo that history in a modern, algorithmic key.
Ultimately, what is being proposed is not a careful attempt to remove genuinely misleading content from the internet. It is a redefinition of dissent as misinformation. The Centre for Countering Digital Hate, the Global Disinformation Index, and the GO-Science review all treat certain political and economic arguments as though they were mathematical errors. For example, questioning the pace and cost of the net zero transition is framed as a matter of physics rather than a legitimate political and economic judgment. The Global Disinformation Index once pronounced a vaccine statistic “clearly untrue” even though it was drawn from Public Health England’s own data, and then tried to persuade advertisers to defund the publisher. Its chief executive has openly explained that content can be “factually accurate but still extremely harmful.” That is a chilling statement. If content that is true can still be labelled as harmful misinformation, then truth is no longer a defence. Whoever controls the label controls the debate. Britain has not seen anything like this since the Star Chamber decreed in 1637 that nothing could be published unless it had been licensed by the authorities. The licensing regime was eventually overthrown because people understood that freedom to publish means freedom to publish unpleasant, uncomfortable, and controversial things. We are now being asked to hand that power back, wrapped up in the language of safety and trustworthiness. But the evidence does not support the panic, the algorithms are not the radicalising monsters they are made out to be, and the state is not a neutral judge of truth. We should think very carefully before we sleepwalk into a world where the government decides what you are allowed to read, watch, and share. Once that power is built, it will not be easy to dismantle.