AI is shortening the “kill chain” in Iran war, says Canadian legal scholar

TEHRAN- Heidi Matthews, an assistant professor of law at Osgoode Hall Law School, York University, researches international criminal law, the law of war and the legal regulation of political violence. Her work examines the implications of military operations for civilian protection and accountability under international law. As the US-Israeli war against Iran raises renewed questions about the limits of military force and the protection of civilians, Matthews spoke to the Tehran Times about the legality of the war, the lessons of Gaza, the erosion of international legal norms, and the responsibilities of states in preventing civilian harm. Below is the full text of the interview: You call the US-Israeli attack on Iran an unprovoked and unlawful war of aggression. What specifically makes it unlawful under international law? This is a pretty clear answer. Under international law, within the UN Charter framework in place since 1945, states are permitted to resort to the use of force in only two scenarios. The first is where there has been specific authorization under Chapter VII of the United Nations Charter, which we clearly did not have in this case. The second is where there is a credible claim to legitimate self-defense, which we also did not have in this case. So it is very clearly, and unequivocally, an unlawful use of force. This is a position that many scholars of international law have taken, including many dozens in the United States, including the current president of the American Society of International Law. I think it is quite clear that this is an unlawful use of force. The US and Israel have repeatedly framed their actions against Iran in terms of security and self-defense. At what point does self-defense cease to be a legal justification for the use of force? I would say that the language and rhetoric the US has used to justify this war have actually not been rooted in self-defense. They sometimes use those words, but in international law, self-defense means that a state is either responding to an actual armed attack that has already occurred against it, or responding to the real threat of an imminent attack, in the sense that there is an overwhelming threat that leaves no choice of means and no moment for deliberation. That is the legal test. This is sometimes referred to as preemptive self-defense, and it is recognized as legitimate. There is a distinction in the law on the use of force between preemptive self-defense and preventive self-defense. In this case, the United States, in its rhetoric, is drawing on preventive self-defense. That is not a recognized kind of self-defense under international law. All of the American justifications for launching this massive operation last February fell into the category of preventive self-defense. The attacks were allegedly designed to destroy Iran’s future nuclear capabilities, to destroy existing ballistic missile capabilities, and to protect freedom of shipping in the region. Trump also made claims rooted in humanitarian intervention on the basis that the war would assist protesters. None of those arguments reflects a necessity of attack, and the language of preventive self-defense simply isn’t applicable in the law. You have written about the Minab school attack and raised concerns about the lack of political and legal accountability. Who should be held accountable for what happened, and through which legal mechanisms? There is very good reason to believe that this particular attack constitutes a war crime. The most obvious mechanism for accountability for war crimes would of course be the International Criminal Court. That is not currently an option here, because neither the United States nor Iran is a party to the Rome Statute. This means the Court has no jurisdiction over this specific attack, or over other alleged war crimes in the context of this war. There are a few ways this could change. The Security Council is entitled, should it so choose, to refer a situation to the ICC, but given the US veto power this is highly unlikely. Another option is that Iran could accept the jurisdiction of the Court through an ad hoc declaration under the Statute for a defined period of time. That period could go back, for example, to the start of the war on February 28th, or perhaps even earlier, to the attacks in the summer of 2025. That would of course mean that not only US officials but also Iranian officials would become vulnerable to investigation and potentially prosecution by the Court. But it is the most obvious answer to the question of accountability. There are some other options that people have put on the table. Some have proposed a special tribunal for Iran, modeled on the special tribunal now being set up in the context of the Council of Europe to deal with Russian aggression in Ukraine. There could also be an international, independent mechanism to comprehensively investigate alleged crimes committed in Iran. This would not allow for direct criminal responsibility, but it would allow for a detailed collection of evidence that could later support trials, including trials in national contexts. We see this now with the mechanism that was set up to investigate crimes in Syria. In terms of political accountability, this really comes down to the willingness of states, specifically US-allied states and Western governments, to call out and name what they are seeing with respect to American action. The most obvious first step is their willingness to label the attacks as unlawful, in other words, as acts of aggression under international criminal law. Until now we have seen a complete political unwillingness to do that. Until Western governments are willing to use the applicable legal frameworks to talk about American conduct in Iran, I think we are not going to see meaningful political accountability regarding what happened, specifically in Minab. You have expressed concern about the use of AI in military attacks. Should the use of AI in military targeting itself face new legal restrictions? This is a really important question, of course, as we see AI growing at breakneck speed. Concerns about the use of artificial intelligence systems in targeting have been at the forefront of the minds of legal scholars who address the regulation of conduct in war. That is partly due to Israel’s use of these targeting systems in Gaza over the last three years, where the evidence points to a high degree of unjustified civilian harm. With the Minab strike, we have further evidence that the use of AI systems, both in gathering and filtering intelligence and in contributing to target identification, is very dangerous. A lot of information, I will say, is still outstanding on this point. The report from the Human Rights Council body investigating the situation in Iran, which we had in September of this year, does not directly talk about the use of artificial intelligence, although reading between the lines it is present. We do have important reporting from places like Bloomberg News, which has talked to many US officials and people familiar with targeting decisions in this war. It gives concerning evidence about the role of artificial intelligence in shortening what is referred to as the ‘kill chain’. What ordinarily would have taken a much longer period for target identification and verification has been compressed into a matter of mere minutes through the use of AI systems, specifically the Maven system. With respect to the Minab school strike, there is good reason to believe that whatever human element was involved in verifying the intelligence and taking the targeting decision may have been influenced by intelligence provided in part by artificial intelligence systems. I worry that the humans who ultimately authorized the strikes may have been predisposed to reflexively trust the reliability of the information being fed to them using AI, especially given that the school had been clearly identifiable as a school for about a decade. Again, much information is still to come, but solid reporting indicates very serious questions. This doesn’t necessarily mean that we need a new international law on artificial intelligence. Existing norms of international humanitarian law can help us address the concrete question at stake, which is the military’s obligation to distinguish between civilians and combatants when targeting. When making those decisions, a military must take all feasible precautions to ensure that the target is actually a military target. For reasons including the use of AI in this case, it looks very much as though sufficient precautions were not taken. In that context, we can still pinpoint the responsibility of military leaders, including military superiors who knew or should have known that their subordinates were not in a position to take adequate precautions. So I am not sure we need new law. I think we can interpret existing law in ways that will get us the results we need around these questions. You have connected the war against Iran to what you call the “Gaza playbook.” What aspects of the conduct seen in Gaza are being repeated in Iran, and which concern you most? The Gaza playbook contains a few elements. I have done some work on this, and other scholars have as well. We have been looking at how certain patterns of conduct that have emerged over the last three years in Israel’s attacks on Gaza have been reproduced not only in Iran but also in Lebanon. I would highlight a few elements. The first is the intentional undermining of the legal framework governing when states may lawfully use force at all. When we talk about this war being a war of aggression, and about states being very unwilling to name it as such, I think we are seeing an undermining of that framework. About 25 years ago, around the time when the United States invaded Iraq, there was a lot of discussion about this particular framework. The United States tried, unsuccessfully, to argue that there was a legitimate case for intervention based on self-defense there. Again, today we just don’t have serious arguments around self-defense, and I think that in part this has been done intentionally in order to undermine the legal framework on the use of force. The second element I would refer to as the normalization of indiscriminate attacks. These are attacks that either do not adequately distinguish between civilians and combatants, or are undertaken in circumstances where the harm to civilians is disproportionate in relation to the military advantage expected to be obtained. Indiscriminate attacks have been a hallmark of Israel’s military operations in Gaza, with the UN commission of inquiry finding that Israeli authorities have given the “blanket authorization” to target civilians “widely and indiscriminately.” We can see attacks by the United States that directly target civilian infrastructure, for example bridges, ports, energy facilities, electrical infrastructure and civilian airports. There are certainly other examples. These attacks on purported ‘dual use’ infrastructure may be disproportionate in that their destruction may not offer a sufficiently important military advantage to justify the attack. I would also say that Iran has conducted attacks in a similarly indiscriminate fashion. We have seen Iranian strikes on power plants, desalination plants and other kinds of civilian infrastructure such as airports and energy facilities, and sometimes residential areas have been targeted. Those strikes would also be unlawful under international humanitarian law, so it is important to point that out as well. Thirdly, I would say there are attacks against civilian objects that are specially protected under the law, such as health care facilities. We see that being normalized across all of these conflicts. Fourth, we see the weaponization of humanitarian assistance. This is particularly acute in Gaza, of course, but it also has elements in the other conflicts. In particular, attacks are undertaken in the knowledge that they will have very serious impacts on the civilian population’s ability to receive the essentials of life, including access to electricity and fuel, for example. Lastly, I would focus on the quite open use of what many have characterized as genocidal discourse by American and Israeli leaders. This