In the first three posts of this series, we covered distinction (Article 48 AP I) and precautions in attack (Article 57 AP I), including cyber operations. This post builds on those foundations. It explains how planners and commanders avoid indiscriminate attacks in practice. It then provides two examples of unlawful indiscriminate attacks in Article 51(5)(a) and (b) AP I.
This matters for operational leaders and defense decision makers because Article 51 AP I indiscriminate attack rules turn civilian protection into control requirements. Military leaders must control what they aim at. Military leaders must also control what their attacks can be expected to do.
Estimated reading time: 8 minutes
1. What are “indiscriminate attacks”?
Article 51 AP I helps military leaders deal with three different targeting challenges. Readers often confuse them, but they require different legal checks.
(i) Intentional attacks on civilians or civilian objects
These attacks deliberately target civilians as such, or civilian objects as such. The law prohibits them. Article 51(2) AP I makes that rule explicit.
(ii) Indiscriminate attacks
These attacks do not necessarily intend to hit civilians. The failure is different. The attacker selects or executes means or methods that cannot distinguish between civilians and military objectives. The attacker designs an attack in a way that makes it strike civilians and military objectives without distinction.
(iii) Disproportionate attacks
These attacks aim at a legitimate military objective. But the commander expects incidental civilian harm that is excessive compared to the concrete and direct military advantage anticipated. Article 51 AP I treats this as a form of indiscriminate attack.
What “indiscriminate” means in operational terms
An attack can be unlawful even where the commander honestly tries to strike a military objective. The law also requires more than correct intent. The commander must ensure that the attack can be directed in the required way. The commander must ensure that the effects of the attack can remain limited as the law requires. That is why Article 51(4) AP I focuses on direction and effects.
Operational test question:
Can you control this attack so it does not strike civilians without distinction?
2. Article 51(4) AP I: The gatekeeper rule
Article 51(4) AP I acts as the central legal filter for targeting decisions. It tells leaders when an attack becomes indiscriminate.
The rule
“Indiscriminate attacks are prohibited. Indiscriminate attacks are:
(a) those which are not directed at a specific military objective;
(b) those which employ a method or means of combat which cannot be directed at a specific military objective; or
(c) those which employ a method or means of combat the effects of which cannot be limited as required by this Protocol;
and consequently, in each such case, are of a nature to strike military objectives and civilians or civilian objects without distinction.”
The provision identifies three situations which we address in more detail below. In any one of them the attack is unlawful.
2.1 Article 51(4)(a) AP I: Attacks not directed at a specific military objective
What the rule prohibits
Article 51 (4) a AP I prohibits attacks “which are not directed at a specific military objective.”
Plain-language meaning
This is “aimless” or “area” violence. The attacker engages no specific military objective. UK doctrine illustrates this as firing off an automatic weapon at no particular target (see UK Joint Service Manual (JSP 383), para. 5.23 and subparas. 5.23.1–5.23.2).
Operational test question:
Can you identify the objective you plan to attack, as a specific military objective?
Example
A soldier fires an automatic weapon “at no particular target.” That action provides a clear textbook illustration of an attack not directed at a specific military objective.
2.2 Article 51(4)(b) AP I: Means or methods not capable of being directed
What the rule prohibits
Attacks that “employ a method or means of combat which cannot be directed at a specific military objective.”
Plain language meaning
Some weapons or delivery methods lack the accuracy required for the setting. If you cannot direct the attack onto the objective, you must not use that weapon or method in that environment.
Operational test question:
Given the operational environment, can you direct this weapon or method onto the specific military objective?
Example
UK doctrine cites the SCUD missile attacks during the 1991 Gulf conflict as an illustration of an inaccurate weapon used against populated areas (see UK Joint Service Manual (JSP 383), para. 5.23 and subparas. 5.23.1–5.23.2)
2.3 Article 51(4)(c) AP I: Effects cannot be limited as required
What the rule prohibits
Attacks that “employ a method or means of combat the effects of which cannot be limited as required by this Protocol.”
Plain language meaning
This rule focuses on uncontrollable effects. Even if you can “aim” at a military objective, the effects may escape your control in time or space. The ICRC Commentary explains that controllability depends on context. Means and methods can be lawful in one setting and unlawful in another (see ICRC Commentary on AP I, discussion of Article 51(4)(c), para. 1961). The Commentary also describes situations where effects become completely out of control.
UK doctrine explains subparagraph (c) as covering situations where the attacker cannot control effects, including dangerous forces (see UK Joint Service Manual (JSP 383), para. 5.23 and subparas. 5.23.1–5.23.2).
Operational test question:
Can you limit the expected effects of the attack, in time and space, as the law requires?
Example
The ICRC Commentary provides a classic illustration: The choice to use a 10-ton bomb to destroy a single building. This instructive example explains why that selection makes extensive damage to neighboring buildings inevitable. It then shows that a less powerful weapon would suffice (see ICRC Commentary on AP I, discussion of Article 51(4)(c), paras. 1961–1965).
3. Article 51(5): Two examples of indiscriminate attacks
3.1 Article 51(5)(a) AP I: Area bombardment
What the rule prohibits
“An attack by bombardment by any methods or means which treats as a single military objective a number of clearly separated and distinct military objectives located in a city, town, village or other area containing a similar concentration of civilians or civilian objects.”
The elements, in decision logic
An area bombardment exists when all of these are true:
- You identify multiple military objectives.
- You see that these objectives are clearly separated and distinct.
- You recognize a concentration of civilians or civilian objects in the area.
- You nevertheless treat them as one single objective.
The ICRC Commentary links this prohibition to carpet and saturation bombing. It explains that the drafters had historical practice in mind. The Commentary stresses that areas between objectives are not themselves military objectives (see ICRC Commentary on AP I, para. 1968).
Operational test question:
Are you treating several distinct military objectives as one “target area” inside a civilian concentration?
Practical meaning
If you can attack objectives separately, you must attack them separately in populated areas.
3.2 Article 51(5)(b) AP I: Lack of proportionality as an indiscriminate attack
Nations now integrate civilian-harm mitigation and IHL alignment into acquisition rules (e.g., U.S. CHMR-AP, emerging NATO standards). Procurement authorities increasingly ask:
What the rule prohibits
Article 51(5)(b) AP I prohibits attacks “which may be expected to cause incidental loss of civilian life (and or) injury to civilians, damage to civilian objects or a combination thereof, which would be excessive in relation to the concrete and direct military advantage anticipated.”
Plain language meaning
This is an ex ante assessment. Commanders must judge expected incidental harm before they strike. They must rely on information reasonably available at the time. Even a very important military advantage cannot justify foreseeable excessive civilian harm.
Operational test question:
Based on current information, is expected incidental civilian harm excessive compared to the concrete and direct military advantage anticipated?
This post introduces the rule but defers the detailed balancing method to the next installment.
4. AI supports human decision making
AI does not change Article 51 AP I. Human legal duties remain human legal duties. But disciplined AI decision support can help commanders and system designers reduce the risk of indiscriminate effects.
Where AI can reduce Article 51 risk
The CNA study identifies concrete AI-supported functions that can reduce civilian harm risk factors, including:
- AI can alert operators to transient civilian presence near the target area.
- AI can detect new changes that undermine earlier civilian harm estimates.
- AI can alert users to misassociation of tracked entities over time.
- AI can recognize and flag protected symbols for human review.
These tools can strengthen Article 51 control requirements. They can help reduce mistaken assumptions. They can help monitor effects that might otherwise escape control.
UNIDIR also emphasizes the value of data-driven processes. It shows how analysts can use better data to update collateral damage estimates. It explains how planners can adjust weaponeering variables accordingly (see UNIDIR, Leveraging Data to Reduce Civilian Harm During Military Operations in Populated Areas, recommended mitigation practices, p. 31).
The constraints that keep AI helpful, not harmful
Current AI support is only as safe as the workflow around it. Military leaders must manage several risks:
- Commanders must require human verification for key judgments.
- Commanders must build auditability into data and model outputs.
- Designers must plan for challenge and review, not rubber-stamping.
- Operators must keep AI inside a targeting process that demands evidence.
Operational instruction:
Commanders should treat AI as a decision-support tool. They must not treat AI as an independent rule.
If commanders ignore those constraints, current AI can accelerate errors rather than reduce them.
Key takeaways
- Article 51 AP I indiscriminate attacks rules require control over direction and effects.
- Article 51(4) AP I acts as a gatekeeper with three “gates.” Any one makes an attack unlawful.
- Article 51(5)AP I provides two examples for indiscriminate attacks: area bombardment and the lack of proportionality.
- AI can support compliance by improving detection of civilians, change monitoring, and collateral damage estimation (see CNA, Leveraging AI to Mitigate Civilian Harm, conclusions and recommended functions list, pp. 52–53.
- Commanders must verify AI outputs and avoid automation bias
Next up in this blog series: Proportionality in practice, including decision logic for “excessive” judgments.
If you would like to discuss particular aspects of designing military AI systems for IHL compliance or if you have any questions on specific LOAC principles, get in touch with me by email or just give me a call.