Imagine a rally in Dhaka in early 2026. By evening, a drone video is circulating on Facebook, claiming to show the ‘biggest turnout in a decade’. A small, well-regarded fact-checking outlet, Shotter Alo, jumps on it the same evening. Their team compares the new footage against old rally clips, spots a mismatch in how the shadows fall, and publishes a verdict: ‘Misleading – Old Footage Reused’. Within hours, that label spreads even faster than the original claim did. Two days later, three independent journalists release footage from angles unseen before. The rally was real – same day, same size, exactly as claimed. What looked like reused footage was just distortion from the drone’s lens. Shotter Alo quietly updates its verdict to ‘Correction Issued’, but almost no one sees it. Supporters of the rally say the outlet was working against them all along. Critics of the rally say it folded under pressure. Neither side trusts the correction. Both sides just go back to believing what they believed before.
This scenario is imaginary, but every fact-checker in Bangladesh has lived a version of it. The claim, the rush, the wrong call, the quiet correction nobody reads – this pattern repeats across newsrooms and fact-checking desks with a regularity that should worry all of us. The mistake here did not start with malice, but with speed. The team felt the pressure every fact-checking desk feels: publish the verdict before the claim spreads unchecked, or lose the moment entirely.
A single piece of evidence rarely tells the whole story. A shadow angle, a missing watermark, a suspicious upload time – these details raise questions, but they don’t answer them on their own. The team treated one red flag as a final verdict. That single choice cost them their credibility for months, maybe longer.
Fact-checking exists to reduce harm, not to referee political debate. Meta’s own fact-checking guidelines make this priority explicit. They ask fact-checkers to focus on claims about elections, claims that target an ethnic or religious group, claims about medicine, and claims that risk major financial loss to ordinary people. These claims can trigger real-world damage: a riot, a communal clash, a health scare, a financial scam.
A doctored image once helped spark attacks on Buddhist temples in Cox’s Bazar. A false rumour once triggered violence against Hindu families in Brahmanbaria. Every fact-checker should remember these events, because they explain why the job exists. It exists to catch a claim before it becomes a match thrown into dry grass, not to settle who scored better points in a talk-show debate.
Somewhere along the way, a chunk of the fact-checking community drifted from this purpose. Instead of chasing claims that threaten communal peace or public health, some desks chase claims that generate clicks. A politician’s exaggerated rally number gets rated within hours. A viral health scam that could poison a village waits days for attention.
This drift creates tokenism. Fact-checking becomes a badge a desk wears to look neutral, while its actual output leans toward whichever claims draw the most traffic or outrage. Rating a politician’s word choice or a party’s turnout estimate might win attention, but it rarely prevents harm. Meta’s guidelines actually exclude a politician’s own statements from fact-checking in most cases, precisely because political speech deserves open scrutiny through debate and journalism, not a verdict stamp from a third party.
Netizens have noticed this drift, and they have started keeping score. When a desk rates one party’s claim as false within a day but sits on the other party’s claim for a week, people stop seeing a neutral referee. The public wants an answer within the hour. A responsible fact-checker needs enough time to confirm the answer is right. These two demands pull against each other every single day, and the desk that gives in to the first demand pays for it later, the way Shotter Alo did in our imaginary rally story.
A credible newsroom rarely leads with breaking news it hasn’t confirmed twice. A credible fact-checking desk should hold itself to the same standard. One wrong verdict, corrected quietly a week later, damages public trust far more than ten claims left unrated for an extra day. A useful fact-checking desk asks a simple question before it touches any claim: does this claim, if it spreads unchecked, hurt real people?
Opinion columns, political endorsements, and a politician’s own quoted words sit outside the fact-checker’s job description for good reason – democracies need open political speech, not a scorekeeper standing between citizens and their leaders.
A fact-checker should not need to know which party a person supports before deciding whether their claim is true. That identity only becomes relevant when it directly shapes a harmful claim. Outside that narrow case, a fact-checker who starts profiling people by political leaning has already stopped fact-checking and started something else.
Fact-checkers who keep this line in view protect their own credibility. Fact-checkers who blur this line, chasing political drama for attention, hand their harshest critics all the ammunition they need.
Netizens deserve fact-checkers who resist the pull of political theatre. Fact-checking organisations deserve a public that judges them by their consistency, not by which side benefits from today’s verdict. Both sides need to remember the same lesson our imaginary Shotter Alo learned the hard way: a fast wrong answer travels further than a slow right one, but only the right one protects the community it claims to serve.
The views expressed in this article are solely those of the author
The writer is a Research Coordinator, FactWatch
