How do you spot fake news?

Effective false information is rarely a big lie: far more often it's a displaced truth. The mechanisms behind it, the markers to look for and the reflexes that defuse them.

Updated on 16 July 2026

The phrase "fake news" has become so political that it barely means anything: it serves equally to name a manufactured lie and to disqualify an article someone dislikes. Two more precise words are better. Misinformation is a falsehood spread in good faith, by someone who believes it. Disinformation is manufactured and spread knowingly, to produce an effect. The content can be identical; the intent changes everything, including what to do about it.

The founding misunderstanding

People picture false information as an entirely invented article on a crude website. That form exists, but it's the least dangerous: it's spotted fast. The content that actually deceives is built out of true elements — an authentic photo, a real quote, an official figure — assembled so as to produce a conclusion the facts don't support. That's why "it's verifiable" is no defence: every brick can be true and the wall still be false.

The most common techniques

Context shifting: a real image or video, but from another place or another year. Nothing is doctored, everything is false.
The truncated quote: the sentence exists, but the half that qualified it is gone.
The figure with no denominator: an impressive raw number, presented without the thing you'd need to compare it against for it to mean anything.
Impersonation by appearance: a site, logo or account imitating a known outlet down to one letter.
Recycled satire: a text published as a joke, reshared away from its home site and read straight.

Why it works on everyone

Believing false information isn't a failure of intelligence, it's how attention normally works. We accept more readily what confirms what we already think, what makes us angry, and what people close to us passed on. The most-shared content is therefore selected — by us, not by a conspiracy — for its emotional charge rather than its accuracy. Knowing this doesn't make you immune; it only lets you notice the moment you're about to share without having read.

An unpleasant thing has to be added: speed always works against verification. False information needs nobody to circulate, whereas debunking it means finding the source, reading it, understanding it and explaining it. The correction therefore arrives structurally later, and reaches fewer people, because it's less spectacular than what it corrects. It's an asymmetry no platform has solved, and it explains why the only genuinely effective moment to act is the one before sharing.

What to do when you meet one

Don't share it, not even to denounce it: an outraged share is still a share, and the algorithm doesn't tell the difference.
Look for the fact, not the debunk: if the event were real, it would have left other, independent traces.
If someone close to you passed it on, reply privately and with the source, not publicly and with sarcasm. Public humiliation hardens positions; it has never persuaded anyone.
Report it where the platform allows, and move on. This fight isn't won retail.

Frequently asked questions

Is false information necessarily manipulation?

No, and that's an important distinction. An honest mistake, a mix-up between two people with the same name or a loose translation produce falsehood with no intent at all. A serious outlet makes them too — the difference is that it corrects and flags them instead of letting them live on.

How do you recognise an AI-generated image?

The classic visual tells — hands, unreadable text, incoherent reflections — are getting less reliable, and betting on them is a poor medium-term wager. The robust question remains provenance: where does this image come from, who published it first, has an agency or an outlet distributed it independently? A photo of a real event almost always has cousins shot from another angle.

Is false information a new phenomenon?

No. Rumour, the forged document and propaganda are as old as writing. What changed is speed and cost: spreading a fake needs neither a printing press nor a distribution network, and a piece of content can circle the world before a newsroom has finished its first phone call.