As militaries worldwide race to integrate artificial intelligence into their air defense systems, Aviv Bar Zohar, a former Israeli Air Force colonel and researcher on autonomous defense systems, argues that the transformation will not come in a single leap but through a gradual process of building trust between humans and machines.

“Artificial intelligence will change the future of air defence and aerial warfare, but not all at once, rather in a gradual process in which trust will be built between the human and the machine,” Bar Zohar said. “It will be possible to estimate with a high degree of certainty how the machine will operate so that one can rely on it, just as people rely on autonomous driving or current driver assistance systems.”

He envisions a shift in aerial combat “in such a way that uncrewed vehicles will fly alongside crewed platforms, where the uncrewed vehicles are the majority, take the risks, and operate under the guidance of the human in the operator loop, while making significant use of artificial intelligence, enabling a small human crew to operate many assets.”

When it comes to air defense, however, Bar Zohar believes change will take longer. “The basis of air defence will continue to be detection, and detection depends less on artificial intelligence,” he explained. “In later stages of identification and prosecution, artificial intelligence will assist the human in the loop. Still, the decision, at least in the coming years, will be made by the human in the operator loop.”

“Detection capabilities depend on sensors—radars, ELINT, COMINT, acoustics—where AI will assist but will not be the single most significant factor,” he said. “AI will help create a more reliable air picture from multiple sensors, but this will not be magic; it will be a gradual process of improving detection, required as the number of targets increases, as we see in Russia and Ukraine.”

Bar Zohar cautioned against overestimating the maturity of fully autonomous drones and swarm attacks. “Contrary to what some might think, today there is almost no use of autonomous drones and swarm attacks. The technology is still not mature,” he said. Even in Israel, “the vast majority of systems are controlled by humans, even if AI recommends the target.”

Looking ahead, he foresees AI taking on greater roles in decision-making but stresses the need for ethical and legal oversight. “In the not distant future AI will be able not only to recommend a target but also to select it and decide on its strike without human involvement,” he said. “States will have to decide how to treat the matter… and ensure significant human involvement in the lethal action, all while complying with international humanitarian law.”

For Bar Zohar, artificial intelligence should be viewed as an aid to human judgment rather than a replacement. “A robot does not want revenge, it always obeys the firing orders, it makes fewer mistakes, it does not get tired,” he noted. “Therefore the combination of a robot on the battlefield together with a human in a place that is not threatened will, in my view, lead to a better outcome that will save lives on both sides.”

He describes this as a “shared team model of joint human-machine control,” one that requires investment in training, education, and trust-building. “Modern armies are already working on building trust between human and machine both in offensive and defensive models and must invest a lot of thought, training, instruction and resources to do this in the right way.”

On the issue of electronic warfare, Bar Zohar emphasized that AI can help overcome jamming and communication disruptions. “In the age of AI better solutions to jamming will be found… from microprocessors that will do the thinking on the asset at the edge up to field recognition mechanisms,” he said. “It will be necessary to limit the decision range of the autonomous tool so that there is a protection mechanism and the autonomous tool will not make large mistakes but will attack clear military legitimate targets.”

He sees AI’s most immediate advantage in intelligence processing. “AI is good at processing information, improving existing information and making it accessible to the human at the edge,” he said. “The human must be the one making the decisions, while the machine will do it at a stage when there is sufficient trust between human and machine.”

Trust, he stressed, “is one of the central topics and requires focus. Because you need to trust the lethal tool or the tool that offers the intelligence insight. If you do not trust what the tool proposes you will not use it.”

Bar Zohar remains confident that robust safeguards will accompany the spread of AI across defense systems. “The risks exist of course as they exist with government databases, hospitals, power stations. Modern organisations know how to take steps to protect information,” he said.

He pointed to Israel’s Iron Dome as an example of deliberate human oversight. “Even today systems like Iron Dome can operate fully autonomously… We choose not to do that because the human in the operator loop constitutes a safety belt that should not and must not be relinquished.”

Ultimately, Bar Zohar believes the future of air defense lies in cooperation, not replacement. “The human will always have an advantage over the machine not in calculation speed but in deeper understanding,” he said. “I always see significant human involvement in the operator loop and sufficient human judgement applied, which together with autonomous tools and significant AI capabilities will greatly improve accuracy, the ability to make correct decisions under short time constraints, and reduce killing and suffering on both sides.”

Aviv Bar Zohar is a retired Colonel in the Israeli Air Force with over 35 years of experience in air defense and autonomy integration. He serves as Senior Advisor to Israel’s National Drone Initiative and as Acting Director of the UAS Experimental Test Center. Bar Zohar is completing his PhD in law at Reichman University, focusing on the legal and ethical challenges of lethal autonomous weapon systems, and frequently speaks at military briefings and academic forums on counter-drone operations, airspace protection, and human–machine teaming in modern warfare.