Feasible precautions in attack are the legal duties a military commander owes to civilians before, during, and after a strike under international humanitarian law. The rule comes from Article 57 of Additional Protocol I to the 1949 Geneva Conventions, which requires anyone planning or ordering an attack to take “all feasible precautions” to avoid or minimize civilian harm.1International Committee of the Red Cross. Protocol Additional to the Geneva Conventions of 12 August 1949 – Article 57 The same duty binds forces as customary international law in both international and non-international armed conflicts, so it reaches states that never ratified the Protocol.2International Committee of the Red Cross. Customary IHL – Rule 15 Principle of Precautions in Attack In practice, the rule breaks down into a series of concrete steps: verify the target, choose weapons and methods carefully, warn civilians when circumstances permit, weigh anticipated harm against military advantage, and stop the attack if the situation changes.
What “Feasible” Means
Everything in the framework turns on this one word. Feasible does not mean every conceivable precaution. It means those that are practically possible given all the circumstances at the time, including both humanitarian concerns and military realities.2International Committee of the Red Cross. Customary IHL – Rule 15 Principle of Precautions in Attack The standard is what a reasonable person would have done with the information available at that moment. Germany made this explicit when ratifying Additional Protocol I: decisions are judged on what was known then, not on what became clear afterward.3International Criminal Tribunal for the former Yugoslavia. Judicial Supplement – The Prosecutor v Stanislav Galic
That framing cuts both ways. If a step is practically achievable and would reduce civilian casualties, it becomes binding. Urgency of the mission, quality of intelligence, risk to friendly forces, and the actual protective value of the measure all feed into the assessment. A precaution that is technically possible but offers no real safety benefit may not qualify. One that is inconvenient but would save lives almost certainly does. There is no fixed financial or logistical threshold; the analysis is contextual. Some states have clarified that the duty falls on those with both the authority and practical ability to act, typically at the battalion level and above.2International Committee of the Red Cross. Customary IHL – Rule 15 Principle of Precautions in Attack A squad leader reacting in seconds is not held to the same standard as a general planning an operation days out.
Verifying the Target
The first duty is to do everything feasible to verify that the object of attack is actually a military objective and not a civilian person or protected building.1International Committee of the Red Cross. Protocol Additional to the Geneva Conventions of 12 August 1949 – Article 57 Verification is continuous, not a one-time check. Satellite imagery, signals intelligence, and human sources on the ground all feed into it, and new information can change the picture.
When genuine doubt exists about whether a person is a civilian, the law resolves the question in their favor. Additional Protocol I states directly that “in case of doubt whether a person is a civilian, that person shall be considered to be a civilian.” The presumption forces attacking forces to hold fire rather than fire first and investigate later.
Modern militaries also work from no-strike lists that catalog locations protected under international law or the rules of engagement for the operation. Diplomatic facilities, religious and cultural sites, hospitals, schools, refugee camps, prisoner-of-war camps, and dams whose destruction could flood civilian areas typically appear on those lists.4Department of Defense. CJCSI 3160.01 No-Strike and the Collateral Damage Estimation Methodology A listed site loses its protection only if it is actively being used for a military purpose, the adversary has been warned, and the warning has gone unheeded.
Choosing Weapons and Tactics
Once the target is verified, the attacking force must select the weapons and methods least likely to harm civilians while still achieving the objective. Article 57 requires all feasible precautions in “the choice of means and methods of attack” to minimize incidental civilian harm.1International Committee of the Red Cross. Protocol Additional to the Geneva Conventions of 12 August 1949 – Article 57
“Means” refers to the weapons themselves. If a precision-guided munition can hit the target with a smaller blast radius than an unguided bomb, and both are available and suitable, the law pushes toward the precision option. “Methods” covers the tactical choices around delivery: approach angle, timing, delivery heading, aim point. Timing a strike for when fewer civilians are likely to be present, or adjusting a trajectory to direct blast effects away from populated areas, are the kinds of decisions this obligation demands.
The U.S. military formalizes this analysis through a five-level Collateral Damage Estimation process that begins with validating the target and identifying nearby protected sites, then progressively refines the weapon, warhead, and fuze combination to reduce collateral risk.5Department of Defense. CJCSI 3160.01A No-Strike and the Collateral Damage Estimation Methodology At higher levels, analysts apply specific mitigations such as delayed fuzing to bury the blast, restricted attack headings, and adjusted aim points. When collateral damage still appears unavoidable, the process requires calculating a casualty estimate and comparing it against a threshold set in the rules of engagement.
Autonomous and semi-autonomous weapons introduce their own precautionary demands. DoD Directive 3000.09 requires that these systems be designed so commanders and operators can exercise appropriate human judgment over the use of force.6Department of Defense. Autonomy in Weapon Systems – DoD Directive 3000.09 Interfaces must show operators clearly what the system will do on its own and what needs human input. Systems using artificial intelligence must be tested rigorously and monitored continuously to make sure changes in the operating environment have not degraded their ability to avoid unintended engagements.
Warning Civilians in Advance
Article 57(2)(c) requires that “effective advance warning shall be given of attacks which may affect the civilian population, unless circumstances do not permit.”1International Committee of the Red Cross. Protocol Additional to the Geneva Conventions of 12 August 1949 – Article 57 The same rule exists in customary law. A warning is “effective” only if it reaches the people at risk and gives them enough time to get to safety. Radio broadcasts, leaflets, and direct electronic messages are common. In some recent conflicts, forces have used small warning munitions on rooftops shortly before a full strike, a technique sometimes called roof knocking, to signal that occupants should evacuate.
The “circumstances do not permit” exception is narrow. It exists mainly for cases where operational success depends on surprise, and it cannot become a routine justification. If a warning is withheld, planners should expect to explain why in any later legal review. Effectiveness also depends on whether the warning is understandable to the population receiving it, which is why U.S. Army doctrine requires Civil Affairs forces to develop the language skills and cultural knowledge needed for genuine communication with civilians.
Weighing Harm and Halting the Attack
Before launching, the commander must assess whether the expected civilian harm would be excessive compared to the concrete and direct military advantage anticipated.1International Committee of the Red Cross. Protocol Additional to the Geneva Conventions of 12 August 1949 – Article 57 If it would be, the attack must not proceed. The law uses “expected” and “anticipated” deliberately; the decision is judged on information available at the time it was made, not on how things turned out.
The duty continues once the attack begins. Article 57(2)(b) requires halting a strike if it becomes apparent that the target is not actually a military objective, or that civilian casualties would be excessive relative to the military advantage.1International Committee of the Red Cross. Protocol Additional to the Geneva Conventions of 12 August 1949 – Article 57 If civilians unexpectedly enter the strike zone, the calculus shifts and the original justification may no longer hold. Real-time monitoring of the engagement is essential to meeting this obligation.
The International Criminal Tribunal for the former Yugoslavia set out the practical standard in the Galić case: whether “a reasonably well-informed person in the circumstances of the actual perpetrator, making reasonable use of the information available to him or her, could have expected excessive civilian casualties to result from the attack.”3International Criminal Tribunal for the former Yugoslavia. Judicial Supplement – The Prosecutor v Stanislav Galic Many violations are judged on that reasonable-person basis rather than on what the commander subjectively knew.
What the Defending Side Owes
These duties do not sit only on the attacker. Article 58 of Additional Protocol I places corresponding obligations on the party controlling territory where military objectives are located.7Office of the United Nations High Commissioner for Human Rights. Protocol Additional to the Geneva Conventions of 12 August 1949 To the maximum extent feasible, the defender must evacuate civilians from areas near military objectives, avoid placing military assets within or near densely populated areas, and take other measures to protect civilians from the effects of operations.
The defender’s breach does not relieve the attacker of its own precautionary duties, but it does factor into the feasibility assessment. A defender who deliberately co-locates military equipment with civilians is engineering a situation where the attacking force faces a much harder proportionality question.
How U.S. Forces Apply These Rules
The United States has not ratified Additional Protocol I, but it treats many of the Protocol’s provisions, including the Article 57 precautions, as reflecting customary international law that binds U.S. forces. The Department of Defense Law of War Manual adopts the same definition of feasible: precautions that are “practicable or practically possible, taking into account all circumstances ruling at the time, including humanitarian and military considerations.”8Department of Defense. DoD Law of War Manual (Updated July 2023) The manual acknowledges that the “fog of war” shapes what is feasible in any given moment; feasibility is a balancing act under pressure, not a fixed checklist.
Precautionary duties reach individual service members through the Standing Rules of Engagement, which require U.S. forces to comply with the law of armed conflict during all military operations involving armed conflict, regardless of how the conflict is legally classified.9Department of Defense. Standing Rules of Engagement / Standing Rules for the Use of Force for US Forces (CJCSI 3121.01B) The rules also build in de-escalation: when time and circumstances allow, forces demonstrating hostile intent should be warned and given the chance to withdraw before force is used. Commanders can tighten these rules for a specific mission with supplemental measures that restrict certain weapons or add proportionality constraints.
What Happens When Precautions Fail
Failure to take feasible precautions can trigger criminal liability at multiple levels. The Rome Statute lists as a war crime the intentional launching of an attack in the knowledge that it will cause incidental civilian harm clearly excessive in relation to the military advantage. Individuals convicted of war crimes at the International Criminal Court face imprisonment for up to 30 years, or life imprisonment when the extreme gravity of the crime justifies it.10United Nations. Rome Statute – Part 7 Penalties Proving a violation at the international level requires showing the accused acted “wilfully and in knowledge of circumstances giving rise to the expectation of excessive civilian casualties,” as the Galić trial chamber put it.3International Criminal Tribunal for the former Yugoslavia. Judicial Supplement – The Prosecutor v Stanislav Galic Simple negligence in target verification or weapon selection may not clear that bar, but it can still trigger liability under domestic military law.
Commanders also face liability even when they did not personally order or carry out an unlawful attack. Under command responsibility, a commander is criminally responsible for subordinates’ violations if the commander had effective control, knew or should have known of the crimes, and failed to take all necessary and reasonable measures to prevent or punish the conduct.8Department of Defense. DoD Law of War Manual (Updated July 2023) The “should have known” standard means commanders cannot avoid liability by refusing to look at what their troops are doing.
Within the U.S. military, service members who ignore precautionary requirements built into orders or rules of engagement can be prosecuted under the Uniform Code of Military Justice. The UCMJ does not name feasible precautions as a specific offense, but violating a lawful order that incorporates those requirements falls under Article 92, and unlawful killings resulting from a failure to take precautions can be charged under Articles 118 and 119 for murder and manslaughter.11Office of the Law Revision Counsel. 10 USC Subchapter X – Punitive Articles Article 134 reaches other neglect prejudicial to good order and discipline, including failures to adhere to the law of armed conflict.
Separately, the Federal War Crimes Act criminalizes grave breaches of the Geneva Conventions and certain other law-of-war violations committed by or against U.S. nationals.12Office of the Law Revision Counsel. 18 USC 2441 – War Crimes The statute expressly excludes lawful collateral damage; death or injury incident to a lawful attack does not qualify. The gap between a lawful attack with incidental harm and an unlawful attack with excessive harm is exactly where the feasible precautions framework operates.