AI Is Now Selecting Military Targets. The World Is Only Beginning to Reckon With What That Means.

The strikes began on February 28. Thousands of U.S. and Israeli munitions have since hit targets across Iran in what military analysts are already describing as the most AI-assisted bombing campaign in history. And while the geopolitical dimensions of the conflict dominate the headlines, a quieter, more consequential story is unfolding in the background: artificial intelligence is now helping decide what gets hit — and what doesn't.

Experts believe AI-driven targeting tools have played a significant role in guiding the ongoing U.S. and Israeli strikes on Iran, though the exact nature and extent of their use have not been officially confirmed by either government. What is confirmed is that the technology is deeply embedded across modern military operations — and that its integration is accelerating faster than the legal and ethical frameworks designed to govern it.

From Logistics to the Kill Chain

Artificial intelligence's role in modern warfare extends well beyond targeting. Today, military-grade AI is deployed across logistics, reconnaissance, electronic warfare, cybersecurity, and information operations. As Laure de Roucy-Rochegonde of the French think tank IFRI noted, "almost any military function can be boosted with AI."

But it is the technology's role in compressing the "kill chain" — the sequence of decisions between identifying a target and striking it — that has drawn the sharpest attention. AI algorithms can sift through massive volumes of data, including satellite imagery, radar signals, drone footage, and real-time video, at speeds no human analyst can match. The result, proponents argue, is faster, more comprehensive threat identification.

U.S. forces are currently deploying the Maven Smart System (MSS), a Palantir-developed tool designed to identify and prioritize potential targets. This week, The Washington Post reported that Anthropic's Claude generative AI model has been integrated into Maven to enhance its detection and simulation capabilities — a development that neither Anthropic nor Palantir has publicly addressed.

The Iran Campaign and an Unanswered Strike

The most disturbing incident to emerge from the current conflict underscores just how high the stakes of AI-assisted targeting can be.

Iranian authorities have reported that a school was bombed during the campaign, killing 150 people. Neither the United States nor Israel has acknowledged responsibility for the strike. Independent verification of the incident on the ground has not been possible.

What is known is that the school was located near two facilities controlled by Iran's Islamic Revolutionary Guard Corps (IRGC). Peter Asaro, chair of the International Committee for Robot Arms Control, suggested the strike may represent a case of mistaken AI targeting. "They didn't distinguish it from the military base as they should have — but who are they?" he said. "Human or machine?"

Asaro raised pointed questions about the integrity of the data used in such systems. If AI were involved in selecting the target, the critical question becomes: "How old was the data?" Was the strike the result of a database error — an outdated record that failed to account for the school's presence?

These questions may never be definitively answered. And that ambiguity, critics argue, is itself part of the problem.

Who Is Accountable When AI Gets It Wrong?

The legal and ethical architecture governing warfare was built for a world where human commanders made human decisions. AI doesn't fit cleanly into that framework — and the gaps are becoming impossible to ignore.

"If something does go wrong, then who's responsible?" Asaro asked. It is a question that applies equally to military planners, software developers, and the governments that deploy these systems.

The Israeli military's use of a targeting program called "Lavender" during its campaign in Gaza brought similar questions into public view. Lavender was used to identify targets within a defined margin of error — an acknowledgment, baked into the system's design, that some percentage of its outputs would be wrong. What that margin means in human terms was left largely unaddressed.

French military AI officials have pushed back on the notion that AI systems operate autonomously. Bertrand Rondepierre, head of the French army's AI agency AMIAD, insisted that "military commanders are at the heart of the action and the design of these systems" and that "no military decision-maker would agree to use an AI if he didn't have trust in and control over what it's doing."

But critics note that when AI accelerates decision-making to the point where human review becomes a rubber stamp rather than a genuine check, the distinction between "human-in-the-loop" and "AI-in-control" begins to collapse.

A Moment That Demands Clarity — Not Just Speed

For Hamza Baig, founder of the Automation Institute and CEO of Hexona Systems, the military application of AI raises questions that extend well beyond the battlefield.

"Automation, at its best, amplifies human capability and frees people to focus on what matters most. But the moment we use it to remove human accountability from decisions that end lives, we've crossed a line that technology alone cannot walk back. The question was never whether AI could do this — it's whether it should, and who answers for it when it doesn't."

Baig, who has trained over 30,000 students in automation and built one of the world's most widely licensed automation platforms, argues that the principles governing responsible automation in business — transparency, human oversight, and clear lines of accountability — are the same principles being tested, and in some cases abandoned, in military contexts.

"Speed is not always a virtue," he has noted in discussions about AI deployment. "In business and in war, the cost of a fast wrong decision can far outweigh the benefit of a slow right one."

The Beginning of Something Much Larger

Despite the controversy, military investment in AI is not slowing down. Benjamin Jensen of the Washington-based Center for Strategic and International Studies, who has spent a decade observing AI in military decision-making exercises, described the current moment as just "the beginning."

"The world's armies haven't fundamentally rethought how we plan, how we conduct operations, to take advantage of AI's capabilities," Jensen said. "It's going to take a generation for us to really figure this out."

That timeline may not be reassuring to those caught in the crossfire of the figuring-out process. The infrastructure of AI-assisted warfare is being built in real time in active conflict zones, with immediate and irreversible consequences.

The legal frameworks, international agreements, and accountability mechanisms needed to govern it remain largely aspirational.

What is clear is that the decisions made now — about how much autonomy to grant these systems, what safeguards to require, and who bears responsibility when they fail — will define the rules of warfare for decades to come.

The technology has already arrived. The reckoning is just beginning.