How to check if something online is credible: a five-step approach

Every day something lands in front of you that sounds important and might not be true. A headline that is a little too neat. A video that seems to show something extraordinary. A screenshot in the group chat with no link, no date and no way to tell where it came from.

You do not need to be a journalist to weigh it up. You need five questions, asked in order, and the discipline to stop at the honest answer rather than the satisfying one. Here they are.

1. Where did this come from?

Before you judge the claim, judge the source. Is it a publisher with a track record, or an account that appeared last month? Is there a named author, a date, a page that says who they are? Can you find the original, or only copies of it?

A claim with no traceable origin is not automatically false. It just has not earned your confidence yet, and until it does, it should be treated as a rumour rather than a report.

2. Is anyone else reporting it?

Something genuinely significant shows up in more than one place, reported by outlets that do not share an owner or an agenda. If a major story is real, the major outlets will have it, and usually within hours.

So when one obscure site is alone in reporting something huge, that silence elsewhere is itself evidence. Either the story is not what it appears, or it is so new that nobody has confirmed it yet. In both cases, the honest position is to wait.

3. What is actually being claimed?

A single article usually bundles several claims together with a good deal of opinion, and they do not all stand or fall as one. "Prices rose last year" and "prices rose because of this policy" are two different claims. The first might be solid and the second might be pure guesswork, wrapped in the same confident tone.

Pull them apart. Ask which sentence is the fact and which is the interpretation, and check the fact on its own. It is much easier to weigh one claim than a whole argument.

4. Search the claim in its own words

Copy the specific wording, the most distinctive sentence, and search for it. This one step does more work than any other. It often finds the original source and the context that was cut away, an older version of the story that shows how it has changed, or someone who has already checked it and written up what they found.

If the search turns up nothing at all, that tells you something too.

5. Weigh the evidence, and accept what it can and cannot tell you

Credibility is rarely a yes or a no. A claim can be partly true. It can have a real core with an exaggerated headline on top. It can be genuinely disputed by people who all know the subject well. Where the evidence is thin, the honest answer is that it is thin, not a guess dressed up as a verdict.

This last step is the one most people skip, because "I don't know yet" is an unsatisfying place to stop. It is also the step that separates checking from confirming what you already believed.

When you want the checking done for you

These five questions are the habit. They are also, more or less, what Lucid does when you hand it a link, a screenshot or a video and ask "Is this true?": it looks for the origin, checks for independent reporting, weighs what it finds, and gives you a plain verdict word with the sources beside it. Where the evidence is decisive, so is the verdict. Where it is not, the verdict says so.