When War Bets on Protest: Iran’s Civilians in a Double Bind

[Mojtaba Touiserkani has a Ph.D. in international relations from the University of Tehran]

In early January 2026, Iran went dark. Protests that began as economically driven unrest—sparked by a collapsing rial—spread rapidly beyond Tehran and quickly widened in demands. The state answered with escalating force. On 8 January, authorities imposed a complete nationwide internet shutdown, severing the public’s ability to see, verify, and organise. In the blackout, the crackdown peaked. By any account, thousands were killed.

It was into that crisis that Washington fused hardware with message. President Donald Trump warned that if Iran “violently kills peaceful protesters,” the United States would “come to their rescue,” declaring the U.S. “locked and loaded.” Days later he urged demonstrators to “KEEP PROTESTING – TAKE OVER YOUR INSTITUTIONS” and promised that “HELP IS ON ITS WAY,” while describing a reinforced posture and, later, a “massive armada” headed toward Iran, ready to act “with speed and violence, if necessary.”

Those lines are not just rhetoric. They sketch a theory of change: external force creates openings; civilians convert disruption into political rupture. Public reporting has described U.S. deliberations over options framed not only as deterrence or punishment, but as actions that could embolden protesters, including discussion of targeted strikes on security forces and leaders to inspire renewed mobilisation. Whether or not such plans are executed, the signalling itself alters the civilian risk environment by suggesting that the “decisive step” is expected to come from inside the street.

The warning here is about a protection gap. When coercion is narrated as “rescue,” civilians can be positioned as operational inputs—expected to do what limited strikes cannot reliably deliver: seize facilities, confront coercive organs, and force collapse from within. That positioning pulls two legal regimes onto the same bodies on two different clocks. In real time, international humanitarian law (IHL) sorts people for targeting: civilians are protected unless—and only for such time as—they directly participate in hostilities, a rule reflected across treaty and customary law in both international and non-international armed conflict. Afterward, Iran’s late-September 2025 statute on intensifying punishment for espionage and “cooperation” with hostile states supplies an emergency-tuned attribution architecture that can convert ambiguity into a security file—turning information, connectivity, and even crowd presence in wartime into liability surfaces under accelerated procedure.

A plan can conscript civilians without ever drafting them. It only needs to imply that history will be decided in the space between a strike and a crowd.

When Coercion Recruits the Street

External actors have long tried to influence Iranian politics. What is newer is treating mobilisation as operational input—leverage without occupation. The promise is a shortcut: degrade coercive capacity, and assume the street will deliver what airpower cannot reliably produce. But that shortcut depends on civilians moving into operationally meaningful spaces.

“Take over your institutions” is not a generic call to protest. It cues movement toward the state’s coercive nodes—security facilities, intelligence buildings, detention sites, communications hubs, and command infrastructure. Once crowds move toward those nodes, civilian presence becomes legible through two overlapping lenses at once: the operational lens of military relevance during conflict, and the domestic lens of national security afterward. Outside actors cannot command crowds like units and cannot guarantee institutions will fall on cue. But exposure is incurred immediately—first as bodies near sensitive sites, later as names, devices, and contacts inside a file.

Two Clocks, One Crowd

The first clock runs during hostilities. It is the clock of classification: who remains a civilian, who is directly participating, what may be targeted, and what precautions are feasible. These judgments are made under uncertainty and time pressure, and the hard cases tend to cluster where civilians and military objectives are proximate.

The second clock runs after. It reconstructs intent from archives—uploads, contact graphs, device histories, and metadata—often through specialised security channels that move quickly and narrow contestation. A “strikes-plus-street” posture synchronises these clocks on the same bodies: it raises the chance of misreading in real time and then supplies the raw material for reclassification afterward.

The IHL Seam

IHL protects civilians by keeping them out of the target set. Civilians “shall enjoy protection… unless and for such time as they take a direct part in hostilities”—a formulation reflected in Additional Protocol I (Article 51(3)) for international armed conflict, Additional Protocol II (Article 13(3)) for non-international armed conflict, and customary IHL. Loss of protection, when it occurs, is act-specific and temporary; “for such time as” prevents permanent transformation of civilians into targetable persons.

The hard cases live at the seam where conduct looks like “help” rather than combat. The ICRC’s Interpretive Guidance describes direct participation as a demanding cumulative test: an act must be likely to harm the adversary’s military operations or capacity (threshold of harm), must directly cause that harm (direct causation), and must be specifically designed to support one party to the conflict against another (belligerent nexus). Most protest, most speech, and most documentation do not meet this threshold.

Some conduct is comparatively clear. Relaying coordinates for imminent strikes, physically guiding forces, sabotaging infrastructure actively supporting ongoing operations, or cyber operations that directly disrupt critical military systems can cross into direct participation when linked to the conflict’s conduct and effects. The more pervasive risk in digitised crises is classificatory drift and error. Live streams and rapid uploads can reveal routes, timings, access points, and patterns of life; when layered with other intelligence, that visibility can function as target acquisition. A civilian may think they are witnessing; an operations cell may see spotting.

Proximity creates a second pressure point. When rhetoric cues crowds toward coercive organs, bodies and devices cluster around sensitive nodes. Proximity does not, by itself, remove civilian protection. But it increases the risk of over-reading in the environments where IHL’s safeguards are hardest to operationalise: dense urban settings, mixed crowds, fragmented intelligence, and compressed strike windows.

The law preserves safeguards in uncertainty. In case of doubt whether a person is a civilian, that person must be considered a civilian (Additional Protocol I, Article 50(1)). Parties must take all feasible precautions in attack to minimise incidental harm (Additional Protocol I, Article 57; and customary IHL). A strategy that hints civilians should “press the advantage” works against these safeguards by encouraging precisely the visibility and proximity that increase misreading risks. And even when a civilian truly crosses the direct-participation line, IHL does not convert that person into a combatant with combatant privileges. When hostilities end, the forum shifts from battlefield to courtroom—but the exposure does not disappear. It changes shape.

Iran’s Late-September Statute: Attribution Built for Emergencies

If the first clock classifies under fire, the second reconstructs from files. Iran’s late-September 2025 law on intensifying punishment for “espionage” and “cooperation” with hostile states is best read as an emergency attribution architecture: it hard-codes hostility, defines culpable conduct in functional security terms, widens “assistance” across multiple domains, and then accelerates procedure once a case is treated as a security file.

The statute begins by treating the United States and Israel as “hostile,” while assigning the Supreme National Security Council the authority to designate—and to lift—hostility for other states, regimes, and groups. It defines “operational action” and “information activity” by capability to endanger national security, ranging from killing and destruction to creating public fear and disrupting communications networks, information systems, and critical infrastructure. It then extends culpability well beyond classic intelligence-gathering: “assistance” can be framed across security, military, economic, financial, and technological domains, while the heaviest exposure attaches to weaponised toolsets—arms, drones, cyber disruption, and sabotage-adjacent conduct—once intent is characterised as cooperation with hostile actors.

The law also pulls media circulation and the street itself into the same security grammar. It criminalises a wide band of political, cultural, media, and “propaganda” activity framed as producing fear or harming national security, including sending images or information to foreign outlets or to designated hostile/adversarial pages—categories that authorities can publicly list and periodically update. Crucially, it separately criminalises unlawful marches and gatherings “in time of war,” turning crowd presence into a liability surface once the state asserts a wartime frame.

Connectivity becomes another evidentiary surface through the prohibition on unauthorised satellite internet devices—explicitly including systems like Starlink—covering personal possession and use, with sharper exposure for facilitation, distribution, installation, and supply, and further escalation where intent is characterised as hostile. Finally, the emergency gear is built in: penalty aggravation for offences committed during war or a declared security or military situation, and out-of-turn processing through designated Revolutionary Court branches with compressed timelines and expansive space for pre-trial detention. Speed here is not merely efficiency; it is deterrence-by-design—especially when attribution is assembled from traces rather than tested through full adversarial contestation.

This is not theoretical. During the January 2026 unrest, judiciary officials publicly indicated that “recent unrest” files were being pursued through this statutory framework, citing its core provisions as a working basis for charges. The second clock is already running.

One Act, Two Vulnerabilities

Put the two clocks together and the bind becomes clear. During hostilities, ordinary civilian behaviour can be pulled toward the participation seam or misread under operational pressure. Afterward, the same behaviour can be narrated as “cooperation” through an emergency law built to convert ambiguity into culpability.

Consider a banal act: filming the aftermath of violence near a sensitive facility to warn neighbours to stay away. In real time, the clip may reveal access routes or security-force movement. Later, the same file can become part of an attribution story: where it was sent, which channel is treated as hostile, what else was on the device, which contacts appear around it, and what intent is inferred from the digital trail. The civilian’s purpose may be protective; the interpretive environment can still make the act legible in both directions.

That is why the protection gap cannot be solved by generic advice to “be careful.” It is produced by design, and it is deepened by any strategy that treats civilians as the conversion mechanism between military disruption and political outcome.

Keeping Civilians out of the Plan

Civilian protection cannot be bolted onto strategy after the fact; it has to be built in upstream.

First, separate military pressure from civilian mobilisation. If coercion is threatened or used, success must not depend on crowds confronting coercive organs, seizing facilities, or feeding the battlespace with real-time information.

Second, treat rescue-framed coercion as escalation, not reassurance. It cues risk-taking and strengthens domestic narratives of orchestration—narratives that hard-coded hostility and broad definitions are designed to absorb.

Third, plan for the legal aftermath as part of protection. Assume visibility becomes evidence; assume channels can be reclassified; assume connectivity workarounds can be criminalised; assume procedure can be accelerated.

If this standoff escalates again, the decisive question will not be whether Washington can hurt Tehran—it can. The decisive question is whether any campaign can remain bounded without converting civilians into instruments—first of coercion, then of retribution. In a system designed to punish legibility, the most protective choice is to refuse the wager and keep civilians out of the plan.