“This is a timely and deeply humanistic work on algorithmic warfare; it exposes the hidden harms inflicted on civilians and places ‘human beings’ at the heart of the debate over the use of artificial intelligence in conflict.” — Jessica Dorsey, Utrecht University Faculty of Law

“This is an insightful and much-needed book that examines how artificial intelligence is reshaping conflict, powerfully placing law, human judgment, and the consequences of algorithms for people at the center of the discussion. It is well worth reading.” — Paddy Walker, J. Leon Group

“Compelling yet heartbreaking! Moffett reveals the real-world impact of algorithmic warfare on civilians and makes a convincing case for a human-centered approach.” — Mark Lattimer, Ceasefire Centre for Civilian Rights

“This book poses a crucial question: Does the use of artificial intelligence in war exacerbate or mitigate harm to civilians? At a time when tens of thousands have already fallen victim to ‘code,’ this is undoubtedly a timely and necessary call for a human-centered approach to warfare.” — Damian Copeland, Article 36

★ A groundbreaking empirical work by Professor Luke Moffett of Queen’s University Belfast, a leading international authority who has conducted field research in more than ten conflict zones, including Colombia, Iraq, and Ukraine. This book brings “algorithmic warfare” back to reality, transcending binary oppositions through policy analysis that directly addresses the hidden harm AI inflicts on civilians in modern warfare.
★ From autonomous drones and intelligent targeting systems to lethal autonomous weapons, the book draws on real-world evidence from Ukraine, Gaza, and other regions to reveal how algorithmic logic creates a dangerous disconnect between decision-makers and victims, offering a critical “human-centered” governance framework for military compliance, legal accountability, and civilian protection.
★ For readers concerned with technological ethics, international security, and human rights in times of war, this is an essential reference for clarifying the role of algorithms in warfare and ensuring the legitimacy of military operations.

By giving voice to the victims themselves, this work illuminates the human impact of war.
Artificial intelligence and algorithms are now driving the evolution of modern warfare, shaping surveillance and targeting. Despite their impressive precision, these systems often obscure severe and hidden civilian harm.
War algorithms transcend the binary of prohibition versus promotion, instead striving to protect humanity through better design and governance.
Drawing on testimonies from conflicts around the world, including Ukraine and Gaza, this book reveals how algorithmic logic creates a dangerous distance between decision-makers and those who bear the consequences. Ultimately, it argues that as AI continues to reshape the battlefield, human agency and human-centered approaches are essential for ensuring accountability, strengthening civilian protection, and upholding the legitimacy of military operations.

Luke Moffett
Luke Moffett is a Professor of Human Rights and International Humanitarian Law at Queen’s University Belfast, and a leading scholar in the fields of international human rights law, international humanitarian law (the law of war), and transitional justice. He has conducted extensive fieldwork in more than ten countries worldwide that have experienced or are currently undergoing armed conflict—including Colombia, Uganda, Iraq, Nepal, Cambodia, and the Great Lakes region—carrying out hundreds of interviews to explore how best to address the consequences of armed conflict. Together with victims, governments, and armed groups, he has worked on the role of reparations in addressing past abuses, as well as on the place of law in limiting warfare and safeguarding the rights of victims.
He completed his LLB, LLM in Human Rights Law, and PhD in Law at Queen’s University Belfast, and has been teaching there since 2013, having since been promoted to full professor.
He frequently contributes to podcasts, blogs, and major news outlets on issues related to the laws of war and their impact on civilians. His publications include “War and Reparations” (Oxford University Press, 2023), “Victim Justice under the International Criminal Court” (Routledge, 2014), and “The Handbook of Transitional Justice Research” (Elgar, second edition, 2023).

Introduction

1. Machine Decoding: Understanding Algorithms, Artificial Intelligence, and Machine Learning
2. Regulating Warfare to Safeguard Humanity
3. The Rise of Drones, Algorithms, and the Cosmos in Warfare
4. Piercing the Fog of War: Algorithms, Data, and Information
5. Artificial Intelligence in Decision-Making and Targeting in Warfare
6. Regulating Lethal Autonomous Weapons Systems
7. The "Triple Black Box": Accountability, Responsibility, and Transparency in Artificial Intelligence
8. Compliance or Catastrophe in the Future? The Good, the Bad, and the Ugly of Algorithms in Warfare

“Faster, Not Smarter: The Rise and Risks of Military Artificial Intelligence,” September 2026
— A podcast article by Luke Moffett explores how the growing use of artificial intelligence on the battlefield is reshaping life-and-death decisions, while also risking the dehumanization of civilians by reducing them to mere data points.

On July 6, 2026, Ukrainian civilians were fleeing to escape yet another Russian drone attack. In Zaporizhzhia, they had no way of knowing that this swarm of drones had slipped beyond the control of its operators; an AI algorithm had taken over, using a newly developed system to guide the drones toward their targets.

The drones had been trained on data identifying fuel tanks found near gas stations. However, one drone miscalculated its trajectory and detonated as Tetiana Bubynets (19), Oleksiy Svirin (41), and Roman Karpii (48) were running toward a shelter, killing all three instantly. This marked the first known instance of a lethal autonomous weapon system—or “killer robot”—being used against civilians.

Even though civilians were not the intended targets, treating every fuel tank as a legitimate target could constitute a war crime. The drone strike came as the United Nations and the International Committee of the Red Cross warned that we are now “dangerously close to crossing a moral red line: machines autonomously attacking humans.” By early September 2026, more than 70 countries had voiced support for advancing negotiations on a treaty to ban killer robots, ahead of the Convention on Certain Conventional Weapons review conference in Geneva this November.

Warnings about the dangers of AI are nothing new. Yet this month, developers and leaders from leading AI companies have repeatedly sounded the alarm about the threats AI poses to humanity.

AI developer Jacob Coxson resigned from Anthropic after previously working at another AI giant, OpenAI (ChatGPT), warning that neither company has taken responsibility… They are embedding themselves into our lives… The people building AI genuinely believe it could kill us all by the end of this decade.

In response, the head of the anthropocentric organization Dario Amodei called for slowing the pace of AI development, establishing independent oversight, and imposing stricter global regulations—views echoed by OpenAI’s Sam Altman and Elon Musk of Grok.

Meanwhile, executives like Meta CEO Mark Zuckerberg and NVIDIA CEO Jensen Huang downplayed such risks, arguing that AI companies should manage risks and build trust.

The threat of AI is no longer confined to science fiction or theory. It is increasingly being integrated into military operations—and neither corporations nor states can be trusted to regulate it effectively on their own.

In Gaza, the Israel Defense Forces employed an AI-assisted targeting system called Lavender, which reportedly generated a database of some 37,000 potential human targets. According to Israeli intelligence sources, the system operated with a “collateral damage” threshold that tolerated significant civilian casualties in exchange for eliminating a single high-value militant.

The United States has likewise incorporated AI into its military operations, using it to identify targets faster, integrate intelligence, and support logistics such as supply chains and vehicle maintenance. Its Maven Smart Systems program alone helped generate over 1,000 targets per day, resulting in more airstrikes within the first 100 hours than the U.S.-led coalition against ISIS carried out in over six months.

Unlike the emerging risks posed by killer robots, the use of AI in strike decision-making has proven to be among the most controversial and deadly for civilians. On February 28, 2026, a girls’ school in Iran’s Minab city was struck, killing more than 156 people, mostly children, just as U.S.-Israeli operations against Iran were beginning. The speed of the AI-driven attack matched the volume of intelligence data available. A subsequent U.S. investigation revealed that the school had been targeted because it remained flagged as a nearby military site dating back years—outdated data, officials noted, rather than a flaw in the technology itself.

A recent UN fact-finding mission to Iran concluded that the targeting of Minab School may constitute a war crime, as the U.S. failed to do enough to verify its status as a legitimate military objective. Using AI in military operations to achieve faster, larger-scale situational awareness and certainty does not obviate the need for adequate human judgment and understanding of legal obligations. AI systems are merely tools, and their impact depends on how humans choose to wield them—both legally and ethically.

This is not the first time the U.S. has deployed AI in military operations. Earlier this year, Anthropic’s Claude model was used during a brief U.S. operation against Venezuela, leading to the capture of former Venezuelan President Nicolás Maduro in Caracas in January 2026. Although Anthropic’s usage policies explicitly prohibit employing Claude to incite violence or develop weapons, this still occurred.

In early September, Anthropic released its latest report on AI misuse, documenting a series of cases from the past year involving the malicious exploitation of its platform. The report aims to highlight the risks of AI and underscore the urgent need for national regulation to prevent even more serious abuses in the future.

Beyond headlines claiming AI will “kill us all” within a decade, the Human + Machine report details how state and private actors are weaponizing large language models—for example, Iran using AI to compile lists of U.S. naval bases in the Middle East, private individuals attempting to make viruses more lethal, and Yemen’s Houthi forces designing and testing advanced missile technologies—all demonstrating AI’s capacity to amplify violence and destructive impact.

Anthropic says it has disrupted every one of the activities described in the report. While other AI providers may face similar challenges, few maintain transparency or monitor misuse.

These developments reveal a broader issue: how the expanding use of AI on the battlefield is reshaping the calculus of life and death. This is a central theme explored in my forthcoming book, “The Algorithms of War.” Yet when AI is deployed in attacks, it risks reducing civilians and other protected persons to data points—calculations subject to rules that are either absent or inadequately enforced.

This November, nations will gather in Geneva to decide whether decades-long negotiations on a treaty banning autonomous weapons will finally become binding law. The question is no longer whether machines should make life-and-death decisions in war. The question is whether states can, before it is too late, prohibit and regulate the clear and present dangers AI poses on the battlefield.