The UK plans a national body to counter disinformation and deepfakes traced to hostile states. The Guardian reports it will detect, attribute and disrupt foreign information attacks, linking intelligence agencies, law enforcement and social media companies.
Information defence in plain terms
“Information defence” is the label governments now use for protecting a country’s public conversation from deliberate manipulation by foreign states. It is narrower than it sounds. It is not concerned with ordinary falsehoods, bad journalism or citizens arguing badly. It targets a specific category: coordinated campaigns run or sponsored by a foreign government to push a narrative, discredit institutions or inflame a domestic dispute, usually while concealing who is behind them.
The Guardian reports that the announcement of a National Centre for Information Defence was made at the United Nations, and that the stated purpose is to detect, attribute and disrupt information attacks by foreign powers, many of which are enabled by artificial intelligence. Those three verbs describe the whole model. Detection means noticing a campaign while it is running. Attribution means establishing, with enough confidence to act, who is running it. Disruption means doing something about it, which can range from public exposure to asking a platform to remove accounts.
The centre’s legal powers, budget, staffing and start date are not established in the reporting available here, and should not be assumed.
Origins of the counter-disinformation state
State-run influence operations are old. The Soviet security services used the term “active measures” for a programme of forgeries, front organisations and planted stories intended to weaken opponents without open conflict. What changed was distribution. Social platforms made it cheap to run thousands of fake personas, target a specific town or demographic, and measure what worked, all from another continent.
European governments responded piecemeal. The European Union set up a strategic communications task force to track and publish examples of pro-Kremlin messaging, and later adopted the phrase “foreign information manipulation and interference” to keep the policy focus on foreign state behaviour rather than on the truth or falsity of individual claims. NATO established a strategic communications centre of excellence in the Baltic region for research and doctrine. In Britain, counter-disinformation work has run through units inside government departments, which attracted criticism from civil liberties groups for opacity about what they monitored.
The newer pressure is synthetic media. Cheap generative tools produce convincing audio of a politician, video of an event that never happened, or hundreds of distinct-sounding comments from a single operator. Audio is the most operationally useful, because it is the hardest for an ordinary listener to check and the easiest to release hours before an election.
Detection, attribution and disruption today
In practice, this work is less about fact-checking and more about pattern analysis. Analysts look for behavioural signals: clusters of accounts created in the same period, posting in a synchronised rhythm, amplifying the same links, or reusing infrastructure such as domains and hosting. Content matters mainly as a fingerprint, because operators reuse phrasing and imagery across networks.
Attribution is the hardest step and the reason a central body is proposed at all. Confident attribution usually requires combining signals that no single organisation holds: technical telemetry from platforms, intelligence from state agencies, and criminal-investigation material from police. The Guardian reports that the new centre is intended to bring exactly these parties together, which is an organisational answer to an evidence problem rather than a technical breakthrough.
Disruption is bounded by law and by the fact that the state does not own the networks involved. Governments can expose a campaign publicly, sanction the companies that run troll farms, prosecute where domestic law applies, or refer accounts to platforms that then enforce their own rules. In the UK, platform obligations sit mainly under online safety legislation and are supervised by the communications regulator, which is separate from any intelligence function.
On deepfakes, detection tools exist but degrade quickly as generators improve. The more durable approach is provenance: cryptographically signing media at the point of capture so that authentic files can be verified, rather than trying to prove a given file is fake.
Common misconceptions
The first is that such a centre decides what is true. The stated model is behavioural — who is coordinating, and are they a foreign state — not editorial. That distinction is what keeps the work outside the domain of speech regulation, and it is also the point at which critics say the boundary is easy to blur in practice.
The second is that these bodies police citizens. The remit described concerns foreign state activity. Whether the boundary holds is a legitimate question, and it is the specific thing civil liberties organisations ask to be written into statute and oversight arrangements rather than left to internal policy.
The third is that disinformation works by making people believe specific lies. Researchers generally find the more achievable goal is corrosive: to make audiences doubt that anything can be verified, so that genuine evidence can be dismissed as fabricated. A related effect, sometimes called the liar’s dividend, is that the mere existence of deepfakes lets people deny authentic recordings.
The fourth is that any of this is settled. There is no agreed method for measuring whether a counter-disinformation campaign changed anyone’s behaviour, and honest practitioners say so.
Where to look next
Read the founding documents when they appear: the terms of reference, the legal basis for any data sharing, and which minister or body answers for the centre in Parliament. Those determine what it can actually do.
Follow parliamentary committees covering national security, intelligence oversight and science and technology, which publish evidence sessions with named officials. Academic groups studying influence operations publish network analyses showing how attribution is actually reasoned. Major platforms publish periodic reports on coordinated inauthentic behaviour they have removed, which are the closest thing to primary data. Digital rights organisations publish the counter-case on scope creep and transparency.
Finally, treat published attributions as claims with confidence levels attached, not verdicts. The useful question about any announcement of this kind is not whether the threat is real, but what the new body can compel, what it can only request, and who can check its work.
Frequently asked questions
What is the National Centre for Information Defence?
According to the Guardian, it is a planned UK national centre intended to detect, attribute and disrupt information attacks by foreign powers, including disinformation and deepfakes associated with hostile states such as Russia. It is described as bringing together intelligence agencies, law enforcement and social media companies. Its powers, funding, location and operating date are not established in the available reporting and should not be assumed from the announcement alone.
Is disinformation different from misinformation?
Yes, and the distinction is about intent. Misinformation is false material spread by people who believe it. Disinformation is false or misleading material spread deliberately to achieve an effect. Policy bodies increasingly prefer a third framing — foreign information manipulation and interference — because it focuses on the coordinated, deceptive behaviour of a foreign actor rather than requiring officials to rule on whether a particular statement is true.
How are deepfakes detected?
Two ways, with different lifespans. Detection classifiers look for artefacts in generated audio, video or images, but they lose accuracy as generation models improve, so they are a race rather than a solution. Provenance is the more durable approach: signing media cryptographically when it is recorded, so verification proves authenticity rather than trying to prove fakery. Neither helps if a convincing clip circulates for a few hours unchallenged.
Does countering disinformation mean censorship?
Not inherently, though the boundary is contested. The described model targets coordinated foreign state activity and covert account networks, not the content of citizens’ opinions. Critics note that the same tooling and relationships with platforms could extend to domestic speech, which is why civil liberties groups press for statutory limits, published remits and independent oversight. Whether those safeguards exist in this case is not known from the announcement.
Which other countries have similar bodies?
Several. The European Union runs strategic communications work tracking foreign manipulation campaigns, NATO maintains a strategic communications centre of excellence, and a number of European states — particularly those bordering Russia — have national units dealing with hostile information activity. Their powers vary widely, from research and public reporting to operational coordination with security services. Direct comparison is difficult because remits and legal bases differ substantially between countries.
Sources and further reading
- The Guardian — the technology desk report announcing the planned National Centre for Information Defence and its stated remit.
- UK Parliament select committee evidence sessions on national security, online safety and disinformation, which publish transcripts and written submissions.
- European Union and NATO strategic communications bodies, which publish open analyses of foreign information manipulation methods.
- Digital rights and civil liberties organisations in the UK, which publish critical assessments of government counter-disinformation units and their transparency.
Surfaced from the rss:guardian_tech signal “new counter-disinformation agency”. AI-assisted draft, editorially reviewed.
