Written by: Long Yue

Source: Wall Street Journal

In the military operations against Iran by the U.S. and Israel, artificial intelligence has been used on a large scale in actual combat for the first time, possibly marking the first "AI war" in history.

According to the latest report from the Wall Street Journal, the involvement of AI tools has significantly improved the efficiency of military operations, from intelligence gathering and target selection to mission planning and damage assessment. In the recent conflict, the U.S.-Israeli coalition utilized AI to process vast amounts of data, quickly lock onto targets, and optimize resource allocation.

However, beyond efficiency, AI has also brought the cost of mistakes to the forefront. In the complex and ever-changing battlefield environment, the limitations of AI have become glaringly apparent. Data aging, system errors, and overreliance on machine decision-making can all lead to catastrophic consequences, potentially resulting in civilian casualties.

Intelligence Processing: From "Endless" to "Retrievable"

Military strikes begin with intelligence, but the bottleneck in traditional processes is "too much to read." Frontline U.S. military officials say that human analysts can typically review only about 4% of the intelligence material.

Colonel Yishai Kohn, who is responsible for planning, economy, and IT at the Israeli Ministry of Defense, stated, "The biggest immediate impact of AI is in intelligence," noting that "many potential missions never happen because there aren't enough personnel to assess critical intelligence."

Reports indicate that the Israeli intelligence agency has long monitored hacked Tehran traffic cameras and eavesdropped on communications of senior officials, increasingly relying on AI to sift through vast amounts of intercepted information for usable clues.

AI's machine vision can quickly identify targets from vast amounts of video and images, even distinguishing specific models of aircraft or vehicles, and can extract and summarize relevant conversations from intercepted voice communications. Matan Goldner, CEO of Israeli software company Conntour, stated, "Intelligence agencies already have vast amounts of video data, and today's AI can help them detect what they need amidst the sea of data."

Acceleration of Mission Planning: From "Weeks" to "Days"

In addition to intelligence processing, AI also shows potential in mission planning and logistics management. Traditional military operation planning requires collaboration among intelligence officials, operational commanders, weapon experts, and logistical personnel, taking weeks to complete. The involvement of AI is expected to shorten this process to just a few days.

In the planning process, any change in detail (such as a shift in target location) can trigger a chain reaction that affects crew scheduling, flight plans, and fuel consumption. In the past, updating these factors was both slow and subjective. Now, AI can instantly process complex interactions and calculate the impact of each change on overall military deployment.

The Pentagon is increasingly using AI to run models and digital war games to optimize target priorities and develop action plans. By processing millions of iterative scenarios, planners can quickly identify the most viable paths to achieving their objectives.

The Double-Edged Sword of Technology: Improved Efficiency and Potential Risks

The application of AI in war still carries significant risks, as warfare is one of the most chaotic and complex areas of human activity.

Jack Shanahan, the first AI chief at the Pentagon and a retired Air Force general, pointed out that a major challenge in building military AI is that much of the data used for training is outdated or unclear. Moreover, errors and inaccuracies in AI systems can lead to fatal consequences on the battlefield. Reports indicate that U.S. military investigators believe that American forces may have caused the deaths of dozens of children at an Iranian girls' school due to intelligence mistakes on the first day of the war.

Even more concerning is the overreliance on AI decision-making. Emelia Probasco, a senior research fellow at Georgetown University's Center for Security and Emerging Technologies, warns that handing decision-making power over to AI "is a serious problem." She emphasizes that appropriate safeguards must be implemented to mitigate risks, while current investment in infrastructure in this field remains insufficient. In war, human judgment is always irreplaceable.