Fact-Checking and Countering Misinformation

Fact-checking is the set of steps you take to confirm that what you are about to publish is true, correctly attributed and presented in its proper context, before publication rather than after. Misleading information is rarely a complete fabrication. Much of it is a real photo from another time and place, a true number with no comparison, or a quote cut out of a longer statement. This guide sets out a practical method for journalists, content writers and anyone running a company's social accounts, because sharing a false story damages a brand's reputation just as much as a reporter's.

The kinds of misinformation you'll run into

Knowing what shape the problem takes gets you halfway to checking it. The most common forms:

• Fabricated content: a story or statement that never happened.
• Genuine content out of context: an old photo presented as today's event, or footage from one country attributed to another. This is among the most common forms and the hardest for untrained eyes, because the image itself is real.
• Manipulated content: a composite image, or audio or video edited in a way that changes its meaning.
• Impersonation: an account posing as a government body, news outlet or public figure, or a website mimicking a well-known newspaper's design and address.
• Misleading numbers: a correct statistic presented to suggest something it doesn't show.
• AI-generated content: images, voices or text that look and sound real.
• Satire taken literally: a joke post that loses its context after being shared a few times.

Specialists distinguish between misinformation, false information people share in good faith, and disinformation, which is created deliberately to deceive. The distinction matters when you write about the phenomenon, but the checking steps are the same.

Before you start: stop and ask four questions

The first step isn't technical, it's mental: stop. If a post makes you angry, scared or very excited, that is a signal it needs more scrutiny, not faster sharing. Then ask:

1. Who is saying this? A known person or organization, or an account you know nothing about?
2. How do they know? Were they there, do they have a document, or are they repeating someone else?
3. What's the evidence? Is there a document, recording or dataset you can examine?
4. What are others saying? Have independent, reliable sources reported the same thing, or does every version lead back to one post?

Then break the claim into small facts you can check one by one. "The ministry canceled exams after the virus spread to 20 schools" contains three separate claims: that a decision exists, what it says, and the number of schools. One can be true and the other two false.

Trace the claim back to its origin

The golden rule of verification: don't settle for the version that reached you. Go upstream to the original.

• A screenshot of a headline: look for the story on the outlet's own website or official accounts. Screenshots are easy to fake, and a story missing from the site is a strong warning sign.
• A quote attributed to someone: find the recording or full transcript and read what came before and after, to make sure the meaning hasn't changed.
• "A study says": find the original study, not the article summarizing it. Check who conducted it, who funded it, the sample size, where it was published and what it actually concluded. The original finding is often much narrower than the headline.
• A decision or law: go to the official gazette, the official website or the verified account of the body concerned.
• A deleted post: you may find a copy in a web archiving service, but treat any archived copy as evidence that needs checking too.

If the trail ends at an anonymous account or at unnamed "informed sources," the claim is unproven, however widely it has spread. Lots of shares do not equal lots of sources.

Lateral reading: check the source before reading it

When you land on an unfamiliar website or account, the natural instinct is to read it top to bottom and judge it by its design and tone. Lateral reading means leaving the page and opening new tabs to find out what others say about the source:

• Who owns or runs the site? Search its name alongside words like "funding," "owner" or "criticism."
• Does the web address match the original outlet exactly, letter for letter, or does it imitate it with an extra character or a different domain ending?
• Is the account verified? When was it created? Has it changed its name recently? What has it posted before?
• Do independent, reliable sources cite it, or does it only cite itself?

A polished design is not evidence of credibility, and neither is a large follower count.

Verifying images

Images are shared more than anything else, and misused more than anything else. The core checks:

• Reverse image search: upload the image to tools such as Google Lens, TinEye or Bing Visual Search and look for its earliest appearance. If it was online years before the event it supposedly shows, you have your answer. Try more than one tool, and try cropping a distinctive part of the image, because results vary.
• Metadata: an original file may carry data such as the camera model, the capture date and sometimes the location. But many social platforms and messaging apps strip or reduce this data on upload, and it can be edited. So missing metadata doesn't prove a fake, and present metadata doesn't prove authenticity. It is one clue among several.
• Context clues: weather, shadows, the language on signs, license plates, architecture, clothing. Does everything fit the claimed place and time? Compare with known photos of the location, maps or street-level imagery.
• AI-generated images: flaws in hands, lettering or reflections can give them away, but those flaws keep getting rarer, and automated detectors aren't conclusive. Rely mainly on provenance: who posted it first, and are there other images of the same event from different angles and independent sources?

Verifying video

The same logic applies to video, with extra steps. Take still frames from distinctive moments and reverse-search them. Listen to the audio: do the language, accent and background sounds fit the location? Find who posted the clip first and when. A familiar pattern in the Gulf and Egypt: old footage of floods, storms or fires resurfaces with every similar event, presented as happening today.

Often the strongest step is contacting the person who filmed it: where did you shoot this, when, and may we use it? Consistent answers that match the other clues are important evidence, and asking permission remains a professional duty, as explained in image rights and licensing.

Numbers and statistics: quick sanity checks

Numbers make writing credible, which is exactly why they get misused. Before publishing any figure, ask about its source and definition: what exactly is being counted, who counted it, and over what period? Then run these checks:

Always test the order of magnitude. If a post says a city of one million people uses one billion water bottles a day, that's a thousand bottles per person per day, which is absurd and suggests someone confused million and billion. Check units and currencies, make sure pie-chart slices add up to 100%, and check whether a chart's vertical axis starts at a value that exaggerates the differences.

Habits that make checking faster and sturdier

• Keep a list of primary sources: official statistics agencies, official gazettes and the verified accounts of the bodies you cover most.
• Document what you did: save dated screenshots, archived copies of pages and a note of what you checked and how. You'll need them if anyone challenges your story.
• Ask people directly: calling the person concerned or an eyewitness settles what tools can't. How to do it well is covered in conducting a journalistic interview.
• Get a second pair of eyes: a colleague reading the piece specifically for errors will catch what you missed.
• Be open about what you haven't confirmed: if a story is important enough to publish before checks are complete, state clearly what is confirmed and what isn't, and attribute everything. If you can't verify a central fact, leaving it out is the safer choice.

A corrections policy: what to do when you get it wrong

Everyone who publishes gets things wrong sometimes. What separates trusted publishers from the rest is how they handle it. A good corrections policy rests on:

• Speed and clarity: correct as soon as the error is confirmed, and add a note to the piece saying what was wrong, what is correct and when it was changed.
• No silent fixes: quietly editing or deleting a piece does more damage to trust than the mistake itself. The exceptions involve people's safety or legal issues, and even then you should say why something was removed.
• Correct in the same channel: if the error went out on social media, post the correction there just as prominently.
• Know the difference: a correction fixes a factual error, a clarification fixes wording that could mislead, and an update adds information that emerged after publication.
• Publish the policy: a short page on your site explaining how readers can report errors and how you deal with them.

These principles are part of the wider framework of professional responsibility covered in journalism and media ethics, and verification's place among all the core writing skills is laid out in the complete guide to journalism and content writing.

Pre-publication checklist

• Did you reach the original source, not just someone passing it on?
• Did you check who the source is by reading laterally?
• Did you reverse-search the image or video and confirm the time and place?
• Did you check every number: its source, definition, units and plausibility?
• Is every quote complete in meaning and in context?
• Is there an independent second source for the central facts?
• Have you documented your verification steps so you can return to them?
• Do you know what you'll do if an error surfaces after publication?