Britain creates a centre to confront networked disinformation

From viral fakes to an information ecosystem

The examples are varied, but their effect is cumulative.

Burnham said the hidden campaign was entering people’s homes, distorting how they viewed their country and community, creating a sense of decline, deepening division and producing despair.

Experts argue that investigating each post separately misses the larger mechanism. Ned Mendes, head of research at the nonprofit Alliance4Europe, calls the hunt for individual examples a losing strategy. Someone can spend a day finding one TikTok video or X post without addressing the problem, he said. The real issue is infrastructure, and the response should focus on the networks distributing the content.

Sophie Williams-Dunning, a cyber-threat analyst at the Royal United Services Institute, describes a complex ecosystem that pushes vulnerable users towards particular actions for different reasons.

That ecosystem is broader than Russia, even though Russian actors have repeatedly amplified divisive content. It also includes:

Williams-Dunning considers Burnham’s announcement useful, though late. She expects it to trigger a dispute over free speech and how far the state should intervene. Hostile networks, she said, will inevitably intensify fears about restrictions on free expression.

The scale behind the stories

Information attacks are not limited to Russia. China and Iran have used AI-generated news presenters to advance their own objectives.

A Social Market Foundation study conducted mainly in February found that more than a quarter of X posts connected to news contained disinformation. One-third of those posts involved anti-immigration or Islamophobic rhetoric.

The British government distinguishes between two related terms. Misinformation is the unintentional spread of false information. Disinformation, in the narrower sense, is the deliberate creation and distribution of lies intended to cause harm or destabilisation. In practice, the two can overlap and produce the same destructive effect.

Labour has become increasingly concerned about Russia’s influence operations in Britain. In May, Keir Starmer’s government said Moscow was waging an information war and had mounted an industrial-scale attack on the information environment of Britain and its allies, seeking to undermine security and democratic systems.

Earlier this year, MPs on the House of Commons Foreign Affairs Committee called for a national centre to counter disinformation. They pointed to Sweden’s Psychological Defence Agency and France’s service for vigilance and protection against foreign digital interference.

The government had already taken several steps. After the Southport riots in 2024, it created the National Online Intelligence Investigations team, which operates under police management and monitors social networks to help local law enforcement deal with threats to public safety.

In May, Britain also sanctioned employees of Social Design Agency and ANO Dialog, Russian organisations that the government says the Kremlin uses for large-scale interference campaigns.

One of those campaigns is Doppelgänger, named for its practice of copying the look of established media outlets, including Le Monde, the Guardian, Der Spiegel and Fox News.

Another involved a surge of posts in 2023 claiming that France was infested with bedbugs and falsely blaming Ukrainian refugees. The disinformation entered the British press, while the French government concluded that Russia-linked social-media accounts had artificially amplified the stories.

The most durable narratives are not invented from nothing

One of the longest-running campaigns claims that London is a dangerous and lawless city. An analysis of Reddit accounts by King’s College London found that the number of such posts rose from 874 in 2008 to 258,444 in 2024.

Robert Topinka, a lecturer in digital media and rhetoric at Birkbeck, University of London, says Russia may not have created such narratives but actively amplified them.

He warns against portraying people as victims of an entirely external manipulator. Russia, he says, usually finds an existing point of tension and then builds more and more content around it. At the centre of these stories is a genuine feeling among some people that Britain has declined.

I think that is the more important diagnosis. A fabricated story can travel because it is false, but it scales because it attaches itself to something people already fear. Removing one video does little when the underlying grievance remains available to anyone willing to package it for outrage.

The announcement is therefore quiet about a difficult question: how will a state distinguish hostile amplification from genuine political anger? A centre focused only on foreign networks may miss the domestic creators, opportunists and audiences that give those narratives reach. A centre that treats every networked grievance as a security operation could reinforce the very distrust the campaign is exploiting.

When chatbots become part of the distribution system

AI adds another layer. Meta, the company owned by Mark Zuckerberg, says AI-generated content has become almost a standard part of disinformation campaigns.

The British thinktank Demos recently described a campaign designed to “poison” chatbots. False information posted online by a discredited Russian organisation was later repeated by chatbot models. One claim said that the bodies of dead Ukrainian soldiers were being processed into meat products.

This changes the target. The campaign is no longer aimed only at persuading people who encounter a post in a feed. It can also seed the material that models later treat as part of the available information environment.

My guess is that the National Centre for Information Resilience will be judged less by how many fakes it identifies than by whether it can map that infrastructure without turning every contested narrative into a government verdict. The harder task is to disrupt the networks while preserving enough trust for people to recognise the difference between protection and control.