In the days before Israeli jet fighters launched the strike that killed Iran’s Supreme Leader Ali Khamenei at his residence, an attack that helped trigger the current regional conflict, Israeli intelligence services had already spent years building a detailed surveillance picture of activity inside Tehran.

According to reporting from The Wall Street Journal, that effort included monitoring hacked traffic cameras and intercepting communications between senior officials, with AI increasingly used to sift through the vast volume of collected data.

AI is now becoming embedded in both the weapons systems and across the intelligence and decision-making processes that precede military action.

Although the U.S. and Israel have declined to detail exactly how AI is being deployed in the current campaign, statements from military leaders and recent reporting suggest the technology is being used primarily as an analytical and operational tool rather than as an autonomous combat system.

From data overload to algorithmic analysis

Modern militaries collect enormous quantities of information from satellites, drones, surveillance systems and communications intercepts. Processing that data quickly enough to inform operational decisions has long been a bottleneck.

AI is increasingly being used to address that challenge. Machine-learning systems can analyse imagery, video and signals intelligence to identify patterns and anomalies that might indicate military activity. Instead of analysts manually reviewing thousands of hours of footage or millions of intercepted messages, algorithms can highlight potentially relevant signals in seconds.

In the context of Iran, Israeli intelligence reportedly relied on AI to process years of intercepted communications and surveillance data. The aim was not simply to collect information but to identify patterns of behaviour that could reveal the location and movements of senior officials.

This shift reflects the growing importance of speed in the intelligence cycle. By accelerating the process of analysing raw data, AI can shorten the time between detection and action.

But the same acceleration also raises concerns among defence analysts. Compressing the timeline for military decisions can increase the risk of misinterpretation or escalation if systems generate flawed or incomplete insights.

An emerging ‘AI-first’ military strategy

The use of AI in the Iran conflict follows years of experimentation by the U.S. Department of Defense and its allies. Washington has invested heavily in military AI programmes since the late 2010s, seeking to maintain a technological edge over rivals including China and Russia.

U.S. Secretary of Defense Pete Hegseth has recently called for the Pentagon to accelerate the adoption of AI in order to create what he described as an “AI-first warfighting force”.

At the same time, the defence department’s reliance on commercial AI companies is becoming increasingly visible. The Pentagon has contracted with OpenAI to deploy its models in classified environments, while also working with other suppliers.

That relationship is not without friction. Hegseth has been engaged in a public dispute with Anthropic, another key AI developer, even as U.S. officials say its AI agent Claude has proven useful in analysing information related to the Iran conflict.

A wider battlefield for AI

The use of AI in operations against Iran is not occurring in isolation. The technology has already been deployed in other recent conflicts.

Ukraine has increasingly relied on AI tools, often developed with support from the United States, to analyse drone imagery, track Russian equipment and coordinate battlefield intelligence. Israel has also used AI-enabled systems in its military operations since 2023.

The conflict with Iran therefore represents part of a broader shift toward what analysts describe as “AI-assisted warfare”, where algorithms augment human decision-making across the intelligence, planning and operational stages of combat.

In the current campaign, the scale of operations underscores the speed at which modern militaries can act once those systems are in place. U.S. officials say more than 3,000 targets in Iran have been struck since the conflict escalated, using platforms ranging from ship-launched attack drones to F-22 fighters and B-2 stealth bombers flying from the United States.

AI does not control those weapons systems, but it increasingly helps determine where and when they are used.

Strategic advantages and new risks

Proponents argue that AI can give militaries a decisive advantage by allowing commanders to make faster and better-informed decisions than their adversaries.

Yet the same technology also introduces uncertainties. Heavy reliance on automated analysis could amplify errors if flawed data or biased models shape operational conclusions. The increasing pace of algorithm-driven decision-making may also compress the window for diplomatic or strategic restraint during crises.

Perhaps most significantly, the Iran conflict shows how AI is becoming embedded in the infrastructure of warfare itself, not just in weapons but in the data pipelines, analytical tools and command systems that underpin modern military operations.

RELATED STORIES

Data centres in the line of fire? What the US–Iran war means for the Cloud

Trump picks OpenAI in Pentagon AI deal, slams Anthropic as ‘disastrous mistake’

Pentagon to weave Musk’s AI into military networks as UK probe looms