In a nondescript building near the Kirya, Israel’s military headquarters in Tel Aviv, a group of soldiers in uniform sit around in front of computers, alongside high-tech employees in civilian dress. At first glance, the room looks like a strange mash-up between a military base and an average startup.
That’s not far from the truth. The office is home to an unusual initiative that involves collaboration between the Israel Defense Forces, high-tech companies, startups and Israeli academics, with the goal of developing an artificial intelligence to decipher and process military intelligence. The initiative is being led by the IDF’s Intelligence Corps and its Administration for the Development of Weapons and Technological Infrastructure, known as Mafat.
>> Read more: For the Israeli army’s tech units, a few rich men | Investigation ■ By providing face-recognition software, Israeli high-tech is on the wrong side of justice | Analysis
Dozens of soldiers, researchers and high-tech employees work on this project, known as Stargate, which is focused on using artificial intelligence to interpret aerial photographs, and on Startrek, focused on non-aerial images. The military has another six such programs working on similar intelligence projects, and hopes ultimately to have 10 in the works.
The goal of all these projects – all named after science fiction movies and series – is to produce intelligence for military operations. “The quantity of data in the world is skyrocketing, artificial intelligence is developing quickly – and the army needs to boost capacity to understand what to do with the intelligence data being collected and how to draw conclusions from it,” says Lt. Col. D., head of artificial intelligence for Military Intelligence’s research and development. “The IDF understood that its current work methods would not address these changes. The key is in partnerships: New relationships with industry and academia.”
- In this secret school, the Israeli army is breeding future cyber warriors
- Israeli weapons makers learn to think more like a startup
- This Israeli face-recognition startup is secretly tracking Palestinians
Two officers show off an aerial photograph of several parked cars. The system automatically identifies the cars to a given degree of certainty. The officers explain that the task is complicated because the photos are sometimes unclear, and also because of the sheer quantity of aerial photographs to interpret.
Until now, the IDF would focus its interpretation efforts on specific locations. But the use of artificial intelligence lets intelligence officers find targeted information. “We’re finding things that an intelligence officer in the field couldn’t dream of being able to ask,” explained D.
One of the challenges is making these tools available to officers in the field. “Once officers would read articles and write documents. Now they need to deal with data, statistics and technological tools. They receive much more training. The officers also instruct us and help us improve the tools,” he said.
Once, such systems were developed exclusively within the IDF, or purchased from external suppliers. There were disadvantages - buying from an external supplier was a slow, complicated process, as developing new capabilities could often take years. Development within the IDF was faster, but the military didn’t always have the necessary manpower.
The IDF hopes that the current method will bring the best of both worlds – the manpower it needs, with experience developed at civilian companies and in academia, alongside the ability to feed the system actual data from the field, while improving the algorithm with the assistance of actual intelligence officers.
“Initially the companies asked, ‘Why should we come to you? We prefer to sell you a product.’ But they discovered we have excellent manpower,” said an officer, Y.
These projects are evidence that the IDF’s impressive intelligence rests partly on the capabilities of the country’s high-tech sector. However, the combination of defense companies, private high-tech companies, academia and the military could lead to problems such as data security and secrecy.
“Everyone who works with us undergoes security clearance, there are data security processes and data can’t be taken outside,” says D.