Israel Has an AI War Machine

In the aftermath of the October 7, 2023, Hamas attack on Israel, the Israel Defense Forces (IDF) faced a challenge. They were incapable of a fast and efficient response. To continue their bombing campaign in Gaza, the IDF turned to a now used tool in modern warfare, Artificial Intelligence. 

Years of development and testing have transformed Israel’s intelligence operations into an AI machine. This is technology that could identify targets at lightning speed and change the pace of the conflict dramatically. But at what cost?

The Emergence of AI in Israel’s Military Strategy

Long before the events of October 7, Israel had been developing a sophisticated AI system that would play a pivotal role in their military operations. This AI-powered system, known as Habsora, or “the Gospel”, was designed to quickly generate targets for Israeli forces in Gaza.

When the IDF’s traditional target bank ran dry, Habsora was ready. It instantly created hundreds of new targets to keep the military’s momentum going. The transformation of Israel’s intelligence unit into an AI testing ground had been years in the making. The IDF’s Unit 8200, one of Israel’s most renowned intelligence divisions, became the center of this AI revolution. It blended machine learning with traditional intelligence gathering practices. 

This shift towards automation and data mining allowed the military to process vast amounts of information from intercepted communications, satellite imagery, and social media posts, enabling soldiers to find patterns and detect threats faster than ever before.

The Role of AI in Combat: Speed, Efficiency, and Controversy

AI’s ability to predict and identify targets in real time was seen as a welcome upgrade by many military strategists. The software didn’t just process raw data: it could identify subtle patterns, such as minute changes in satellite images over years, to locate hidden tunnels or rocket launchers. 

This level of speed and precision was unprecedented, but it raised ethical concerns about the AI’s impact on civilian casualties and the accuracy of intelligence. Critics within the IDF began questioning whether the rapid pace of AI aided decision making compromised the thoroughness and quality of intelligence. 

Some argued that automation in targeting could lead to higher civilian casualties, as AI systems could overlook critical nuances that human analysts might catch. These concerns were particularly significant as reports began emerging of increasing civilian deaths in Gaza, with more than 45,000 casualties reported, including thousands of women and children.

AI Target Spotting

One of the most striking features of the AI system was its ability to generate a large volume of targets, including low-level militants, at an astonishing speed. Using machine learning tools like Lavender, the IDF could assign a probability score to individuals, estimating their likelihood of being involved with militant groups. 

This process, while efficient, raised questions about its accuracy and its potential to misclassify civilians as combatants. While the IDF maintained that their AI systems minimized collateral damage and increased targeting accuracy, some critics contended that these advancements allowed the military to lower the threshold for civilian casualties. 

Historically, Israel’s acceptable ratio of civilian deaths to high-level terrorists was one-to-one, but in the 2023 Gaza conflict, that number had reportedly risen to 15-to-1 for low-level Hamas members, with some sources suggesting even higher numbers for mid- and high-level targets.

A Shift in Intelligence Culture

Under the leadership of Yossi Sariel, the IDF’s intelligence division underwent a profound shift. Once a hub of skilled analysts who carefully considered intelligence reports, Unit 8200 became a technology-driven powerhouse, with engineers taking precedence over language specialists and analysts.            

This dramatic reorganization raised alarms among former commanders, who feared that the focus on AI could undermine the IDF’s traditional strength: human intelligence. The rapid pace of decision-making, driven by machine learning algorithms, meant that critical insights from human analysts were sometimes sidelined. 

And the potential for errors or misunderstandings grew. In particular, AI systems struggled with processing languages and slang used by Hamas, which led to occasional misinterpretations.

Human Control or AI Dominance?

Despite the advancements in AI, the IDF stressed that human oversight remained an integral part of the process. Officers were still required to approve any targeting recommendations generated by the AI systems. 

However, some military insiders raised concerns that AI reliance was blurring the line between human judgment and machine decision-making. This tension was most evident after the October 7 attack. Two former senior commanders argued that the over-reliance on AI had contributed to the intelligence failure. They consequently believe this allowed Hamas to carry out the assault.

The idea of an AI-driven military system, where machines assist in identifying and selecting targets, has sparked broader debates. The role of human decision-making in warfare within the IDF has been questioned. Proponents of AI argue that it improves efficiency and minimizes risks for soldiers. However, critics contend that it could lead to a loss of accountability, particularly when civilian lives are at stake.

The Global Debate Over AI in Warfare

As Israel’s military operations in Gaza continue, the use of AI in warfare is under increasing scrutiny. In addition to internal debates about the ethics of AI aided targeting, Israel’s practices are now facing international attention. Human rights organizations and international bodies have questioned whether the use of AI in warfare meets the standards of international law, particularly in terms of minimizing civilian harm.

The ongoing debate over AI’s role in modern conflict is not unique to Israel. Militaries around the world experiment with AI technologies. But a question hangs in the air: Can machines be trusted to make life-and-death decisions, or is human oversight essential to prevent catastrophic mistakes?

Many seem to lean towards the latter. 

Looking Ahead

Israel’s AI military strategy represents a significant shift in the way modern warfare is conducted. The IDF’s embrace of AI is likely to serve as a model for other nations exploring similar capabilities. However, the rapid pace of these technological advancements raises important ethical and operational questions that must be addressed.

For now, the use of AI in Gaza offers a glimpse into the future of conflict. We could be saddled with a future where machines play an increasingly central role in making human decisions. Whether this trend will lead to safer, more efficient wars or to greater harm remains to be seen.