An illustration of Axon Vision’s EdgeRCWS, a fire control unit embedded with AI. (Axon Vision)
JERUSALEM — The Israeli military is using artificial intelligence in its ground operations in Gaza, the head of an Israeli AI firm said, to aid in situational awareness by sorting through the deluge of data that comes with modern warfare.
“The commander and crew don’t know what to do with all this data,” Roy Riftin, CEO of Axon Vision, told Breaking Defense in a recent interview. “You need something to process it and make it relevant to operational time.”
Axon Vision, founded in 2017, produces tech that can be integrated with ground vehicles, UAVs, including loitering munitions, and maritime platforms. The company says it specializes in “developing cutting-edge technology tailored for the needs of modern warfare, utilizing AI and optical sensors to process vast amounts of data and support real-time decision-making in combat scenarios.”
Its solutions have been integrated into “key platforms of the Israel Defense Forces,” the company says, specifically the army’s Namer and newer Eitan armored personnel carriers, according to Riftin.
RELATED: ‘Useful’ or ‘dangerous’: Pentagon ‘maturity model’ for generative AI coming in June
The tech is increasingly relevant to ground operations, Riftin said, because soldiers in armored vehicles no longer poke their heads out to see around them but rely on external sensors for situational awareness. In addition to the sensors on an APC, that data can now be networked together with optics or sensors on UAVs overhead, providing a more complete picture of the battlefield for the soldiers.
That’s a lot of information to keep up with, and that’s where AI comes in. With so much data to process, AI can help in decision making.
A soldier sitting and watching a screen for hours will become fatigued but “the machine doesn’t get tired,” Riftin said.
The AI can be trained to analyze and detect certain gestures as more threatening than others, for instance. It can differentiate between armored and regular vehicles, he says, noting that “you can teach the AI what you want it to show.”
“It has to do with decision-making and priorities to teach the machines what the priority between different threats and targets,” he said.
RELATED: Israel’s 3 biggest defense companies take stock after 5 months of war
The “biggest challenge,” however, is “not to have false alarms, so soldiers only act with something that is relevant — writing the algorithm and implementing analytics so you don’t have false alarms.” This can be especially helpful in complex urban environments or operating at night, as the IDF often does in Gaza and other fronts.
It’s also key to keep a human in the loop, said Riftin, himself a 30-year IDF veteran and former artillery commander in the IDF.
“What we did with AI software is recognizing the target and threats and push on the display the target and the whole process of slew to cue and tracking the target is all automatic,” eh said. “We don’t take the man out of the loop. The decision making is always done by the man. We have a fire sequence also of a few targets, giving it a priority of targets, to know from first to last in a fire sequence, so you can move from one to the other.”
In an urban environment in which a target may only appear for a few seconds before going back into hiding, Riftin said, accurate analysis of sensor data is key.
RELATED: ‘Not an exaggeration’: Fears rise of Israeli incursion into Lebanon to push back Hezbollah
The IDF uses a plethora of UAVs, from smaller Skylark drones used by artillery groups to Hermes 450 to larger Herons, but does not specify how AI is used with them or on what platforms it may be used. In general, with more sensors on the battlefield processing through AI, targets can be passed from one platform to another, so that an operator can get a broader picture and not lose track of a threat.
In his interview, Riftin said his company is developing tech that uses LiDAR (Light Detection and Ranging) to work with UAVs over areas with vegetation and tree cover.
Elsewhere he said Axon Vision is working on AI for use with loitering munitions. The IDF does not currently use that tech, he said, though the company has “different customers that have this solution.”
But one-off solutions are not the future. “The next level is that today every platform has its specific algorithm and analytics that we use for different electro-optics, but how we use it as a cloud, now it is about building a system rather than just a product,” he said. “As a system it uses next generation use of AI. That’s the idea, it’s a processing challenge.”
Israel’s broader use of AI in its Gaza operations has come under scrutiny, especially in the wake of news reports about an Israel AI program called the Gospel, developed by the Israeli government, that has been used to aid target selection.
Riftin clarified that those systems use AI to crunch big data to produce target banks, which is a broader use than his company’s tech.
“From what we do, it’s not relevant,” he said. “We don’t deal with big data but rather electro-optics that sees the threat and target and is the solution that allows [soldiers] to avoid collateral damage and not enhance it.”
But overall Riftin defended the use of AI in modern combat. The question, he said, “that is not asked what would happen if there was no AI for this operation,” he said.
Without AI and other new technology, soldiers would maneuver with their vehicles and then get to a place and wait.
“Stay in [the] same place for weeks and months; and you’ll do the same work every day and sometimes [become] less alert, and if you have something technological that can give you a better sense of what happens outside, it gives you an advantage,” he told Breaking Defense.
