How AI Is Being Used in War and What That Means for All of Us
I want to be honest about why this topic matters to every person reading this — not just military analysts or defense policy experts. AI tools that started as productivity assistants and research helpers have been integrated into weapons systems that are making life-and-death decisions at speeds humans cannot match. That happened faster than most people realize. And in 2026, we have enough real-world evidence to see both what AI makes possible in warfare and what it costs when it fails.
This post covers what's actually happening — the real use cases, the real incidents, the real investments, and the real questions that no one has answered yet. I've tried to be factually accurate and balanced on a topic that is genuinely disturbing in parts. Please read it that way.
Hours into the same operation, a Tomahawk cruise missile struck the Shajareh Tayyebeh girls' elementary school in Minab, southern Iran — killing at least 168 people, mostly children. A preliminary Pentagon investigation found the strike was likely caused by stale human-curated intelligence fed into the targeting system — not an AI malfunction.
That distinction — human error feeding an AI system versus the AI itself failing — is the central question of military AI. It doesn't make the outcome less devastating. It makes accountability harder to assign.
The Scale of Military AI Investment in 2026
What's happening in warfare with AI is not a future scenario. It's a current reality being funded at enormous scale by multiple governments simultaneously.
In 2026, the US declared it would become an "AI-first" warfighting force. That's not a slogan — it's a procurement and doctrine shift with billions of dollars behind it. And the US is not alone. China, Russia, Israel, India, Turkey, and dozens of other nations are investing heavily in military AI simultaneously, in a dynamic that researchers are calling an AI arms race with no agreed rules.
What AI Is Actually Being Used For in Active Conflicts
Between 2022 and 2026, the conflicts in Gaza, Ukraine, and Iran have served as the first operational tests of AI in war zones, offering early insights into how AI may change warfare forever. Here is what those tests have shown.
AI-Assisted Targeting
AI systems analyze intelligence data, satellite imagery, and sensor feeds to identify and recommend targets. The Maven Smart System compressed targeting timelines from hours to seconds — allowing military forces to strike at unprecedented speed and volume.
Autonomous Drone Warfare
Ukraine has been the proving ground for AI-enabled drone warfare at scale. Small, cheap, AI-guided drones can identify and strike targets with minimal human involvement. The Ukraine war transformed rapidly through innovation in small drone warfare, increasing pressure on militaries worldwide to keep up.
Intelligence and Surveillance
AI processes enormous volumes of satellite imagery, communications intercepts, and sensor data far faster than human analysts. What used to take days of intelligence work now takes minutes — giving commanders a faster picture of battlefield conditions.
Cyber and Electronic Warfare
AI-powered systems identify and exploit vulnerabilities in enemy networks, disrupt communications, and defend against incoming cyberattacks. This form of AI warfare is invisible but pervasive — affecting infrastructure, communications, and financial systems without a single physical strike.
Logistics and Maintenance
AI predicts equipment failures before they happen, optimizes supply chain routing in active conflict zones, and manages the enormous logistical complexity of modern warfare. Less visible than targeting but arguably as strategically important.
Disinformation Warfare
AI-generated deepfakes, fake personas, and coordinated synthetic media campaigns are being used as weapons — designed to confuse enemy populations, undermine trust in institutions, and shape narratives in conflict zones. This overlaps directly with the social media manipulation we covered in a previous post.
The Incidents That Show the Failure Modes
The Minab school strike is not an isolated incident. It is the most documented and deadly example of a pattern that researchers and journalists have been tracking across multiple conflicts.
Investigations by journalists and human rights organizations documented AI-assisted targeting systems being used in Gaza in 2024 and 2025, with credible reports of civilian casualties linked to targeting recommendations from automated systems processing intelligence data. Israeli military officials confirmed the use of AI tools for target generation while maintaining human final authorization in all cases. The debate over what "human authorization" means when AI compresses timelines to seconds remains unresolved.
A 2026 study of 21 simulated nuclear-crisis scenarios across three frontier AI models found that 95% of scenarios included nuclear signalling, 76% featured strategic nuclear threats, and 95% included tactical nuclear use. The researchers emphasized these were simulations — but the finding that frontier AI models consistently escalated toward nuclear options in crisis scenarios is one of the most alarming research results to emerge from military AI research in 2026.
Abdul-Rahman al-Rawi, a 20-year-old student, was the first acknowledged civilian killed by an AI-assisted airstrike in a US strike in Iraq in 2024. The acknowledgment itself was significant — it represents a shift toward greater transparency about AI's role in targeting, while also raising questions about accountability when AI recommendations lead to civilian deaths.
Who Is Developing Military AI — The Global Race
| Country | Status in 2026 | Key Development |
|---|---|---|
| United States | Leading — "AI-first" declared | Maven Smart System, $13.4B autonomous weapons budget, Replicator drone program |
| China | Second — rapid acceleration | AI command system capable of planning large-scale air strike missions involving multiple aircraft and hundreds of targets — unveiled August 2026 |
| Israel | Advanced — combat-tested | Developed AI targeting systems tested in active Gaza and Iran operations — most combat experience of any military |
| Russia | Active — Ukraine proving ground | AI-guided loitering munitions and drone swarms deployed extensively in Ukraine |
| India | Growing fast | Indian Air Force announced AI-based target identification systems for helicopters with precision micro-missiles — August 2026 |
| Turkey | Export leader | Bayraktar drones with AI-assisted targeting exported to 30+ countries — changing warfare economics globally |
The 5 Hardest Ethical Questions With No Easy Answers
Who Is Accountable When AI Makes the Wrong Call?
When an AI targeting system recommends a strike that kills civilians, the accountability chain is genuinely unclear. The soldier who approved it? The commander who authorized the operation? The company that built the AI? The government that deployed it? The Minab school strike investigation concluded the cause was human-provided stale intelligence — but the AI processed that intelligence and generated the recommendation. Where does human responsibility end and AI responsibility begin?
Can AI Make Meaningful "Human Authorization" Possible at Machine Speed?
International humanitarian law requires human authorization for lethal strikes. But when AI compresses targeting timelines to seconds and generates hundreds of recommendations simultaneously, what does meaningful human review actually look like? A human clicking approve on AI recommendations at machine speed is not the same as a human making an independent lethal decision. Most military frameworks have not resolved this distinction.
Does Lowering the Cost of Killing Make Wars More Likely?
AI is compressing the kill chain, lowering the cost of lethality, and embedding private industry into states' military architecture. When a $500 AI-guided drone can destroy a $2 million tank, and when AI systems can plan and execute strikes with minimal human staffing, the cost and risk calculation of initiating conflict changes. Whether that makes war more or less likely is a genuine open question that strategists disagree on.
Should Private Companies Build AI for Weapons?
The Maven Smart System is built by Palantir. Major AI labs have faced internal employee protests over military contracts. The line between civilian AI companies and defense contractors has blurred significantly in 2026 — with major AI labs' models being used in military systems. What obligations do AI companies have to know and control how their models are used?
Can International Law Keep Up With AI Weapons?
A UN resolution passed in December 2025 and a three-day multilateral meeting planned for June 2026 are the first formal steps toward international governance of AI in armed conflict — but a binding framework remains unlikely in the near term. The rate of military AI deployment is far outpacing the rate of international governance. By the time binding rules exist, the technology may be so embedded in military doctrine that effective regulation becomes impossible.
The Positive Side — What AI Could Do for Peace
Honest coverage of AI in warfare requires including the potential benefits — which are real, even if they sit uncomfortably alongside the risks.
AI also has the potential to enhance peace, widening participation in peace processes, offering anticipatory indicators on conflict escalation, and accelerating the recording, sharing and verification of war crimes. AI systems are being used to monitor ceasefire compliance, detect early warning signs of conflict escalation before they become crises, and document war crimes at a scale that would be impossible manually. The same satellite imagery analysis used for military targeting can also monitor mass displacement, document destroyed civilian infrastructure, and provide real-time evidence for international courts. AI is a tool — and like all tools, it can be directed toward prevention and documentation as well as lethality.
What This Means for You — Why Civilians Should Care
If you're not in a conflict zone this might feel distant. It isn't. The AI companies building civilian productivity tools are the same companies whose technology ends up in military systems. The policy decisions being made now about autonomous weapons will define what warfare looks like for the next generation. And the questions about accountability, speed, and human oversight being debated in military AI apply equally to AI systems making consequential decisions in healthcare, criminal justice, and finance.
The Minab school strike is the starkest current example of what happens when AI systems operating at machine speed make recommendations that humans approve faster than human judgment can actually operate. That dynamic — AI speed outpacing human oversight — is not unique to military applications. It's the defining challenge of AI deployment across every high-stakes domain in 2026.
"The question is not whether AI will be used in war. It already is, at scale, with real consequences. The question is whether humans can maintain meaningful control over systems that operate faster than human decision-making can function."
Key Takeaways
This is one of the most important topics I've written about on this blog. I've tried to present the facts as accurately and honestly as I can. If you have corrections or additional context — especially from people with direct knowledge of these conflicts or systems — please share in the comments. Getting this right matters.