At the end of 2023, Israel was aimed at murdering Ibrahim Biari, a high commander of Hamas in the Northern Gaza Strip who helped plan the October 7 Massacres. But Israeli information could not find Mr. Biari, who, according to them, was hidden in the tunnel network under Gaza.
Thus, Israeli officers turned to a new military technology imbued with artificial intelligence, three Israeli and American officials said on events. The technology was developed a decade earlier but had not been used in combat. Finding Mr. Biari gave a new incentive to improve the tool, so the engineers of Israel Unit 8200, the equivalent of the country of the National Security Agency, quickly integrated IA He said, people said.
Shortly after, Israel listened to Mr. Biari’s calls and tested the AI audio tool, which gave an approximate location for where he made his calls. By using this information, Israel ordered air strikes to target the area on October 31, 2023, Kill Mr. Biari. More than 125 civilians also died during the attack, according to Airwars, a London -based conflict instructor.
The audio tool was just an example of how Israel used war in Gaza to quickly test and deploy military technologies supported by AI to a certain extent that had never been seen before, according to interviews with nine officials of the American and Israeli defense, who spoke under anonymity because the work is confidential.
In the past 18 months, Israel has also combined AI with facial recognition software to correspond to partially obscured or injured faces to real identities, turned to AI to compile potential air strike targets, and has created an Arabic AI model to feed an chatbot that could scan and analyze text messages, publications on the social media and other data on the language Arabic, according to the two people who know the programs.
Many of these efforts were a partnership between soldiers enlisted in unit 8200 and reserve soldiers working in technological companies such as Google, Microsoft and Meta, said three people with technology. Unit 8200 has set up what has become known as “The Studio”, an innovation center and a place to match experts with AI projects, people said.
However, even if Israel ran to develop the Arsenal of AI, the deployment of technologies has sometimes led to erroneous identifications and arrests, as well as civil death, said Israeli and American officials. Some officials had trouble with the ethical implications of AI tools, which could cause increased surveillance and other civil murders.
No other nation has been as active as Israel in the experimentation of AI tools in real -time battles, said European and American defense officials, giving an overview of how these technologies can be used in future wars – and how they could also hurt.
“The urgent need to face the accelerated innovation of the crisis, a large part of IT,” said Hadas Lorber, head of the research institute applied to AI responsible for the Holon Institute of Israeli Technology and former senior director of the Israeli National Council. “This led to revolutionary technologies on the battlefield and the advantages that have proven to be essential in combat.”
But technologies “also raise serious ethical questions,” said Ms. Lorber. She warned that AI needed checks and counterweights, adding that humans should make final decisions.
An Israeli army spokesperson said that she could not comment on specific technologies because of their “confidential nature”. Israel “is attached to the legitimate and responsible use of data technology tools,” she said, adding that the army was investigating Mr. Biari’s strike and was not “unable to provide additional information until the investigation is completed”.
Meta and Microsoft refused to comment. Google said he had “employees who reserve themselves in various countries of the world. The work that these employees do as reservists is not connected to Google ”.
Israel previously used conflicts in Gaza and Lebanon to experiment and advance technological tools for its soldiers, such as drones, telephone hacking tools and the Iron dome Defense system, which can help intercept short -range ballistic missiles.
After Hamas launched cross -border attacks in Israel on October 7, 2023, killing more than 1,200 people and taking 250 hostages, AI technologies were quickly authorized for deployment, four Israeli officials said. This led to cooperation between unit 8200 and the reserve soldiers of “The Studio” to quickly develop new AI capacities, they said.
Avi Hasson, CEO of Startup Nation Central, an Israeli non -profit organization that connects investors to businesses, said Meta, Google and Microsoft reservists had become crucial to stimulate innovation in drones and data integration.
“Reservists have brought know-how and access to key technologies that were not available in the army,” he said.
The Israeli army quickly used AI to improve its drone fleet. Aviv Shapira, founder and chief executive officer of Xtend, a software and drone company working with the Israeli army, said that algorithms powered by AI had been used to build drones to lock and follow the targets remotely.
“In the past, the return capacities have relied on the implementation of an image of the target,” he said. “Now, AI can recognize and follow the object itself – whether it is a moving car, or a person – with fatal precision.”
Shapira said his main clients, the Israeli army and the US Department of Defense were aware of the ethical implications of AI in war and discussed responsible use of technology.
A tool developed by “The Studio” was a model of Arab language known as the model of large language, three familiar Israeli officers said with the program. (The large language model was previously reported By more than 972, an Israeli-Palestinian information site.)
The developers had previously struggled to create such a model due to a shortage of Arabic language to form technology. When such data was available, it was mainly in standard written Arabic, which is more formal than the dozens of dialects used in spoken Arabic.
The Israeli army did not have this problem, said the three officers. The country had decades of intercepted text messages, transcribed telephone calls and screened publications of social networks in the spoken Arab dialects. Thus, the Israeli officers created the model of great language during the first months of the war and built a chatbot to execute requests in Arabic. They merged the tool with multimedia databases, allowing analysts to carry out complex research on images and videos, said four Israeli officials.
When Israel murdered Hezbollah Hassan Nasrallah chief in September, the Chatbot analyzed the answers through the Arabic -speaking world, three Israeli officers said. The technology was differentiated between the different dialects in Lebanon to assess the reaction of the public, helping Israel to assess if there was public pressure for a counter-piste.
Sometimes the chatbot could not identify modern argot terms and words that have been translited from English to Arabic, two officers said. This required that Israeli intelligence agents have expertise in various dialects to examine and correct their work, said one of the officers.
The chatbot has also sometimes provided bad answers – for example, the return of pipe photos instead of firearms – two Israeli intelligence officers said. Despite this, the AI tool has considerably accelerated research and analysis, they said.
At the temporary checkpoints set up between the Gaza Strip of the North and South, Israel also began to equip cameras after the October 7 attacks with the possibility of scanning and sending high resolution images of Palestinians to a Facial recognition program supported by AI.
This system also had difficulty identifying people whose faces were obscured. This led to arrests and interrogations of Palestinians who were mistakenly reported by the facial recognition system, two Israeli intelligence officers said.
Israel has also used AI to scrutinize the data Amassaes by intelligence officials on Hamas members. Before the war, Israel built an automatic learning algorithm – called “lavender” – which could quickly sort data to hunt low -level activists. He was trained on a database of confirmed members of Hamas and intended to predict who could be part of the group. Although the predictions of the system are imperfect, Israel used it At the start of the war in Gaza to help choose attack targets.
Few goals have turned out to find and eliminate the senior management of Hamas. Near the list of the list was Mr. Biari, the commander of Hamas who, according to Israeli officials, played a central role in the planning of the attacks of October 7.
The military information of Israel quickly intercepted Mr. Biari’s calls with other members of Hamas, but could not determine its location. They therefore turned to the audio tool supported by AI, which analyzed different sounds, such as sound bombs and air strikes.
After having deduced an approximate location for where Mr. Biari made his calls, Israeli military officials were warned that the region, which included several apartment complexes, was densely populated, two intelligence officers said. An air strike should target several buildings to ensure that Mr. Biari has been murdered, they said. The operation was greenery.
Since then, Israel Intelligence has also used the audio tool alongside cards and photos of the Labyrinth of the Gaza underground tunnel to locate the hostages. Over time, the tool has been refined to find individuals more precisely, two Israeli officers said.