Amnesty's fair words, unfair report

A peculiar irony is found at the heart of the Amnesty International report on the 16 March strike on Kabul's Omid Treatment and Rehabilitation Centre, and it is worth naming before anything else.

Afghanistan, under the Taliban, is a country that has gone out of its way to keep independent eyes out. The UN's own special rapporteur on human rights, Richard Bennett, had been barred from entering the country in 2024, with the Taliban administration accusing him of "spreading propaganda". Women, who make up half the population and are often the only people local custom permits inside hospitals and shelters to treat other women, have been steadily pushed out of public institutions, UNAMA offices included. And yet, for all odd reasons, the official Afghan voices and UNAMA sourcing find their way into Amnesty Report's footnotes without much apparent difficulty. How is it that a country that closes its doors on the world's monitors seems able to get its version of events out, quoting them?

That contradiction alone should have made Amnesty's investigators more cautious than they appear to have been. But it is worth being fair about what strikes like this actually are before getting to what is wrong with how this one was reported.

War, wherever it is fought, leaves scars on the people who never asked to be part of it. Pakistan knows this better than most nations currently living through it. Its own western borders have absorbed years of strikes, raids and cross-border fire, with the dead count filed away under the benign heading of collateral damage, as though a word could soften what a family actually lost. Afghans know it too, arguably worse. During the two decades of the American war against the Taliban, Afghan civilians were killed in numbers that rarely made headlines, dismissed in the same collateral-damage language whenever a wedding party or a farming family was caught in the wrong place. The International Criminal Court itself, in 2020, authorised an investigation into alleged war crimes committed by American forces and the CIA in Afghanistan - a rare moment when a body of that stature was willing to say plainly that even a superpower's conduct in war deserved scrutiny.

That is precisely where this report falls short. Those earlier findings, against the Taliban and against the Americans alike, were built on physical evidence, forensic detail and direct interviews with the people who survived or buried their dead. They named names, dates and units, and let the facts carry the accusation. What Amnesty has produced this time leans far more heavily on testimony it could not independently cross-check on the ground, filtered through a government that has every incentive to inflate a casualty count and none to allow anyone to verify it. A handful of videos and a dozen interviews are asked to stand in for what happened inside a facility said to have held thousands of people that night. That is not the same rigour the organisation has applied elsewhere, and it shows.

None of this suggests civilian harm can be excused as an unavoidable cost of war, or that a stronger state may strike inside a weaker neighbour's territory without its account ever being tested. A hospital's protection under international law does not diminish simply because the source reporting the strike invites suspicion. The real issue is methodological: before asking the world to accept a war-crimes allegation, Amnesty's report needed to meet the same evidentiary bar it holds elsewhere. In its evident haste to validate Kabul's account, the organisation let a grave charge rest on evidence far thinner than it warranted. Such reports never stay confined to the page. They land in a relationship between Pakistan and Afghanistan that is already smouldering, and they do not cool it down. They pour oil on it.