Hanane Djoua
Behind every rule of war is a human being it was written to protect: a child asleep in a family home, a patient in a hospital, a doctor trying to keep that patient alive, or a family deciding whether a warning means they should leave — and, if so, where they can safely go.
International humanitarian law (IHL) has rules for all of them. Civilians may not be deliberately targeted. Parties to a conflict must distinguish between civilians and military objectives, take feasible precautions and ensure that expected civilian harm is not excessive in relation to the anticipated military advantage. Hospitals and medical personnel receive specific protections.
Yet Gaza has forced a difficult question into view: what happens when extensive legal protection exists on paper, civilian harm is documented almost in real time, and civilians still experience destruction on an extraordinary scale?
The question matters beyond Gaza.
Satellites can record destruction. Smartphones can document an attack within minutes. Investigators can identify victims, geolocate footage and reconstruct incidents.
The world has become better at seeing civilian harm. It is much less clear whether it has become better at stopping it.
When the numbers are not the verdict
Airwars published assessments of 606 civilian-harm incidents reported in Gaza between October 7 and 31, 2023. In those incidents, at least 5,139 civilians were reported killed, including at least 1,900 children. Sixty-five incidents involved at least 20 reported civilian deaths.
The dataset covers incidents in which civilian harm was reported; it is not a record of every Israeli strike during that period. And the numbers, however severe, do not by themselves determine whether individual attacks violated IHL.
Professor Janina Dill, Dame Louise Richardson Chair in Global Security at the University of Oxford, told Qatar Tribune that assessing an attack can require information about the target, whether it constituted a military objective, the civilian harm reasonably foreseeable beforehand, the anticipated military advantage and the precautions available.
The IHL Centre makes the same distinction: “The dire consequences of Israel’s attacks may raise concern that it had violated the rules on the conduct of hostilities but do not, in themselves, prove this to be the case.”
This is a central difficulty. Civilians experience what actually happened, while the legality of an attack may depend heavily on what could reasonably have been anticipated before it happened.
Consider the Israeli strike on Bureij refugee camp on October 17, 2023.
Israel announced that it had killed Ayman Nofal, whom it identified as a senior Hamas figure. Airwars subsequently matched Israeli military footage to the location and assessed that at least 92 civilians were killed, including at least 40 children and 28 women. Another militant was also reported killed.
Neither the existence of a military target nor the civilian death toll alone answers the proportionality question. Investigators would need to know what civilian harm was anticipated, what military advantage was expected and what precautions were feasible.
Two weeks later, on October 31, an Israeli strike hit Jabalia refugee camp. Israel said the operation targeted Ibrahim Biari, whom it identified as a Hamas commander. Airwars assessed that between 126 and 143 civilians were killed, including 69 children and 22 women, while 280 civilians were reported injured. It also recorded between 12 and 24 reported militant deaths.
Again, those figures do not by themselves determine legality. But when severe civilian outcomes recur, they raise another question: what should previous harm mean for the decisions that follow?
When yesterday’s harm changes tomorrow’s calculation
“If there are patterns of attacks that harm civilians,” Dill told Qatar Tribune, “over time, it may well become less plausible that this civilian harm was unforeseeable, or unavoidable, or even unintentional.”
IHL generally assesses proportionality ex ante — according to what could reasonably have been expected when an attack was planned. Actual civilian harm does not automatically determine legality.
But the IHL Centre notes that what actually happened may nevertheless indicate what could reasonably have been expected.
If previous attacks repeatedly produce greater civilian harm than anticipated, when should those outcomes change the assumptions behind the next attack?
For civilians, protection may depend not only on the calculation made before an attack, but on whether the system making those calculations learns when reality repeatedly contradicts its expectations.
The law may examine one attack. The civilian lives through everything that came before it.
Much of the law governing hostilities assesses individual attacks. Civilian life does not reset after each one.
A hospital damaged today may have reduced capacity when another neighbourhood is hit tomorrow. Damage to electricity can affect water, sanitation and healthcare. Repeated displacement can push thousands into already overcrowded areas.
The IHL Centre distinguishes between reverberating effects — indirect but foreseeable consequences of a particular attack — and cumulative effects produced by multiple actions over time.
Foreseeable reverberating effects should form part of proportionality assessments. Cumulative harm is more difficult: the Centre notes that there is no specific general rule for assessing a military operation as a whole against aggregate civilian harm.
Yet previous destruction changes the baseline. A healthcare system already severely degraded has less capacity to absorb another disruption.
A joint World Bank, United Nations and European Union assessment published in April 2026 estimated Gaza’s recovery and reconstruction needs at approximately $71.5 billion, based on damage and losses between October 2023 and October 2025.
The legal question may begin with one attack. The civilian experience does not.
When a legal safeguard meets reality
Evacuation warnings reveal the same distance between a legal safeguard and lived protection.
According to the IHL Centre, an effective warning depends on timing, specificity, whether it reaches civilians in understandable terms and whether they have an actual opportunity to protect themselves.
For the civilian receiving it, the questions are practical: Can I leave? Is the route safe? Is the destination safer? What happens if I cannot move?
Crucially, civilians who remain after a warning do not lose their protection. They must still be considered in proportionality assessments, and feasible precautions must still be taken.
The same tension surrounds hospitals. IHL grants medical facilities and personnel special protection, while damage to healthcare can affect patients far beyond the site of an attack through lost electricity, equipment, staff, medicines or access.
Where does protection break down?
The rules governing war are extensive. But Gaza exposes a more difficult question than whether those rules exist.
What turns a legal protection into protection in practice?
Part of the answer lies in decisions made before an attack: the quality of information, how civilian harm is anticipated, the precautions taken and whether previous outcomes change the decisions that follow.
But the protection gap does not end when an attack is over.
Once evidence of civilian harm exists, another question begins: What happens when the law is there, the evidence is there, and protection still fails?
