Operation Epic Fury: The promises and perils of AI warfare

It has been more than a decade since artificial intelligence and military experts were calling AI and autonomous weaponry “the third revolution in warfare,” after gunpowder and nuclear weapons. Whether in the deserts of the Syrian civil war, the plains of the Russia-Ukraine war or the urban landscapes of the conflict between Israel and Hamas, AI has played an increasingly prominent role.  

Now, with the joint U.S.-Israel operation in Iran, codenamed Operation Epic Fury by the U.S. and Operation Roaring Lion by Israel, we are witnessing the next stage of that evolution. The unfolding of this conflict in real time will yield many important signals and implications regarding the future of warfare, as well as the increasingly complex questions about ethics, accountability, human oversight and the potential erosion of restraint in lethal operations.  

Operation Epic Fury was initiated on Feb. 28, with a series of coordinated strikes against key military facilities and members of Iranian leadership. Within just 24 hours, U.S. forces struck nearly 1,000 targets; as of this writing, we’ve hit more than 3,000 targets using stealth aircraft, cruise missiles, precision-guided bombs, rocket artillery and attack drones launched via air, sea and land platforms.  

Amid this flurry of strikes, one of the ways that AI has been used has been the real-time integration of intelligence via satellites, drones, radar and signals data, identifying potential targets and prioritizing them based on their strategic importance, thus enabling simultaneous operations across multiple domains.  

AI has also enabled rapid decision-making beyond what’s possible through human deliberation alone, significantly shortening the duration of the “kill chain,” or the timeline from target detection to engagement. AI-synthesized intelligence and “decision compression” has also enabled these early attacks to target leadership within the Islamic Republic of Iran and suppress their air defenses early on. 

Unmanned aerial systems (UAS) have also been used in addition to, and in conjunction with, AI-assisted target identification and prioritization to dramatically increase strike volume, lower risks for friendly military personnel, and reduce the costs of strikes, since drones are relatively inexpensive compared to fighter jets and missiles.  

At the same time, cyber operations were carried out in a supportive capacity to disrupt Iranian communications and sensor networks in an effort to “disrupt, disorient and confuse the enemy,” according to Joint Chiefs of Staff Chairman Gen. Dan Caine. Iranian traffic cameras hacked by Israel provided real-time intelligence on the whereabouts of key Iranian officials including Ayatollah Ali Khamenei. And we have seen kinetic-cyber merging in instances such as Israel’s strikes on the state-owned TV network, the Islamic Republic of Iran Broadcasting, hijacking the airwaves to broadcast speeches by U.S. President Trump and Israeli Prime Minister Benjamin Netanyahu calling on Iranian citizens to rise up against the regime.  

While each of these elements are not new in themselves, the current operation may be the first instance of all of them being used concurrently. What may have previously taken days or weeks is now being done in hours.  

Another feature of the current conflict is the highly visible degree to which prominent AI companies are closely involved at every step of the kill chain: find, fix, track, target, engage and assess. Anduril’s AI-powered Lattice platform gathers vast amounts of data through a “lattice” of cameras, surveillance drones, sensors and radar. Palantair’s Maven Smart System synthesizes the data from many disparate sources, including Lattice, as well as satellite imagery, signals intelligence and human intelligence, and then analyzes it to identify targets and rank them in order of strategic priority. This in turn is facilitated by Anthropic’s Claude, embedded as a reasoning layer within the Maven system.  

Anduril moreover stays present in the engage stage of the kill chain via its unmanned weapon systems, such as the Bolt-M autonomous attack drones and the Anvil interceptor drones.  

Such close collaborations between the military and AI corporations present some potential complications and conflicts of interest, however. The fact that Anthropic’s Claude is an embedded part of Maven while, at the same time, the company is involved in a conflict with the Pentagon is a vivid example. What happens when the military and government depend on an AI company that they themselves have labeled a supply chain risk? We don’t fully know yet.  

And while this “third revolution” of warfare promises many benefits, they do not come without significant potential tradeoffs or moral and ethical concerns. In many cases, the very advantages themselves are closely intertwined with the risks. 

For instance, while AI systems may enhance speed and tactical advantage, they also add the risk of faster escalation. Bombing quicker than “the speed of thought” is not necessarily ideal if it leads to unintended or undesired (from the human perspective) escalation between major powers. 

There is also the risk of AI misalignment, in which an AI’s goals and behaviors diverge from human goals and values due to faulty training, biased data or any number of other factors. While we do not yet know the factors behind the strike on an Iranian girls’ school that resulted in more than 160 people being killed, a tragedy that is still being investigated, the aforementioned risks of AI-enabled warfare can potentially result in similarly tragic incidents.  

The military AI arms race is another concern. As nations develop more advanced AI, both in general and specifically for military applications, other nations will follow suit. What is used by one nation can conversely be used against it. For example, just as the U.S. and Israel have used cyber operations in conjunction with its kinetic ones, the same tactics and AI-enhanced information warfare and influence operations can be used by Iran against them. We know that Iran has the capability to do this, and there are already signs of heightened cyber threat activity.

Of course, this is only skimming the surface. What will be needed is constant vigilance, oversight and accountability in which humans stay meaningfully involved so that the advancements of AI military technology do not outpace the strategic, ethical and legal risks. We do not yet know what the right balance should look like, but Operation Epic Fury may become the first real-world setting where this balance will be tested.  

AI and autonomous weaponry are indeed the future of warfare. And for better or worse (or both), that future is here and now. The hope is that we can keep the scales tilted toward the better.   

Craig Albert, Ph.D., is a professor of political science at Augusta University, with areas of concentration including international security studies, cybersecurity policy, cyberterrorism and cyberwar.

Copyright 2026 Nexstar Media Inc. All rights reserved. This material may not be published, broadcast, rewritten, or redistributed.

More Opinions - Technology News

White House seeks to circumvent Schumer on potential DHS funding deal

GOP Senate candidate on rising gas prices: ‘Maybe you take one less trip to ...

The Memo: Frustrated Trump struggles against perception that he’s losing ...

Huckabee Sanders says she was asked to leave restaurant because employees felt ...

GOP debate on SAVE America Act morphs into ‘circular firing squad’

Senate GOP plows forward on SAVE America Act amid pressure from right

Jeffries says Democrats will oppose rule governing spy powers bill

Enabling this TSA feature could save you hours in airport security lines

6 governor races shift to Democrats: Election handicapper

Trump’s push to abolish the Education Department reaches student loan ...

Trump’s takeover of DC landmarks reaches legal apex

5 takeaways on Gabbard’s grilling in House worldwide threats hearing

The Memo: Trump fumes as foreign leaders show new willingness to defy him

Thune: Republicans will use SAVE Act in midterms if Democrats don’t get ‘on ...

Why Apple wants you to update your iPhone, iPad now

Selective Service rules: Who would be exempt in event of a draft?

US F-35 makes emergency landing after flying over Iran

Obama’s former DHS secretary calls to end shutdown, pointing to Iran threat

The Hill Podcasts – Morning Report

© The Hill