Israel’s Use of Artificial Intelligence Will Change the Future of War

War is always going to be fought with people and weapons. It is also always going to involve “platforms,” such as tanks and the capabilities they have. It is important to understand that in discussions of the future of warfare the issue is not just about the person or the platform but also tying it all together. At the heart of that effort today are attempts to develop better algorithms and artificial intelligence. This will play an increasing role in war, especially in hi-tech militaries, in the future. As Washington gears up for more contests with near-peer competitors, such as Russia or China, and moves away from the global war on terror, the need for this kind of technology and envisioning the future of warfare increases.

There is no shortage of discussion on the topic of future warfare. An article at the Heritage Foundation in 2018 noted that “to achieve a significant increase in military effectiveness, the new item must be married to an appropriate organization, concept of operations, set of tactics, command and control system, and supporting infrastructure.” Others have questioned whether the future will involve small, smart swarming munitions or exquisite and expensive platforms, such as the F-35 stealth jet fighter and Global Hawk drone. Packing electronic information equipment onto the platforms with the best electro-optics and computers is part of transformation that militaries have gone through in the last decades.

Israel has often been at the forefront of the revolution in military technology because as a small country it needs to stay one step ahead of its enemies and it needs better technology to overcome unique challenges. In the past Israel didn’t have the luxury of time to wait around to win a war, it needed immediate overmatch of its enemies, to dominate the area and the land. To get to this point it developed better defenses, such as Iron Dome to stop smaller missile threats, and developed Trophy to protect tanks. Rafael Advanced Defense Systems, one of Israel’s largest defense companies, was at the center of the effort to invent Iron Dome and Trophy, among other systems. Its origins are in Israel’s government supported defense research and development and it says that is “DNA” today is development of new solutions to defense problems.

One of the Rafael’s development centers is in northern Israel, where the company has a plethora of sites. There are a lot of gadgets that developers are working with. The idea is to experiment with ideas and then plug them into real world programs. For instance, future armored fighting vehicles is an issue that Israel wants to pioneer. That means a vehicle that may have only two men in it, instead of a tank’s usual four-person crew. It could mean more autonomous vehicles, using the kinds of self-driving ideas that civilian car makers are working with. What does a military want out of a future armored vehicle? It needs to have lethality and it needs more sensors to see around it what is happening, to chart and identify threats and feedback information for a commander to decide which “effector” or shooter to use.

Vehicles need to plan the best routes and that means more computer vision solutions, the company’s innovators say. This means translating a real-life battlefield into some kind of augmented reality, such as overlay of the battlefield, more like a computer game in three dimensions than an infantry soldier wandering the streets with the M-16. Another way that computers and algorithms or artificial intelligence can help is in machine learning to help identify vehicles in a complex urban or rural environment. Think of a tank camouflaged or a “technical” truck with a .50 caliber machine gun mounted on the back. Will a person always pick out the right vehicle scanning a city of 10,000 similar trucks that don’t have machine guns on the back? Can a person watching all these trucks follow five of them that appear to be threats? A computer can do it better and faster, sorting the wheat from the chafe and giving the operator and commander the decision on which one to strike or neutralize. To get the computer to do this it needs to think more like a human and it needs to learn and also make the right decisions.

Computers have been making seemingly better decisions than people for a long time. In 1997 IBM chess computer Deep Blue defeated chess champion Garry Kasparov. Fighting a war is like playing chess, but a lot more complex and with real world fatalities.

Rafael says that its developers are learning that they need better autonomous capabilities and better understanding of scenes that sensors and optics see on the battlefield. Rafael’s SPICE family of air to surface missiles, for instance, already use advanced scene-matching algorithms and have independence from GPS. The idea is to increase these capabilities for ground forces, taking the load off operators and leading to minimum collateral damage. In Israel’s recent airstrikes the number of civilians killed has been reduced to almost zero since the 2014 conflict in Gaza. For instance, more than 1,000 airstrikes on Iranian targets in Gaza resulted in few casualties, even to enemy forces. Mostly it was enemy equipment that was apparently damaged. Israel rarely discusses these airstrikes, former Chief of Staff Gadi Eizenkot revealed that there had been more than 1,000 in January 2019, but didn’t say what had been struck. Satellite images of alleged strikes are often published online. The results appear to be pinpoint accuracy.

Rafael, like most military focused companies, needs to operate in the world of algorithms. Last year it unveiled a new Automatic Target Recognition capability for its SPICE 250 munitions. The weapon now uses “deep learning” and artificial intelligence. The pilot or operator is still in the loop, unlike the futuristic movie-induced fear of machines running wild and choosing their own targets. These kinds of systems are most helpful using the new technologies available to militaries, such as drones that can over basically non-stop over a town or village. Rafael has talked up Wide Area Persistent Surveillance capabilities on drones this year, employing a system that can build a 3D model from the surveillance.

Putting all these systems together is the job of the innovators, to figure out how best to integrate them into a kind of ecosystem for vehicles, drones and fighters on the ground. Dr. Ran Gozali, executive vice-president of Rafael and head of the land and naval division, says we are now entering the fourth industrial revolution. “This revolution is autonomous robots, augmented reality, additive manufacturing, the Cloud,” and other elements, he says.

All of these innovations have battlefield applications. Augmented reality is entering the battlefield. In addition, ideas borrowed from the civilian world, such as “smart cities” are being funneled into a “smart battlefield” to connect soldiers and sensors together. The problem for soldiers is that they can’t be holding rifles and using tablet computers all at the same time. They have to be saved from information overload. The same goes for commanders. In past wars, the problem for commanders was often having almost no information, so armies and navies were blind. Admiral Horatio Nelson famously couldn’t catch the French fleet on the way to Egypt in 1798, an effort hampered by lack of visual intelligence. Similarly Robert E. Lee during the Seven Days campaign couldn’t strike at George McClellan’s retreating army at least in part due to lack of good maps of the area. Today’s generals do not have that problem. Civilian mapping technology and communications has in some cases outpaced military abilities in the last decades. The military challenge is to fuse the immense data that is coming in to reduce what appears to be a chaos of information. How do we replicate the human brain, Gozali asks? 

“How do we have the computer give [the commander] the relevant data… we have people that study that, showing that the commander may be under too much pressure and so the computer does more.” This is called the human-machine interface. The issue here is also to help commanders determine, based on the weapon capabilities they have available, which to use. Rafael has a variety of systems now that are relevant for this future battlefield. That includes its BNET software defined radios, Fire Weaver that helps knit together a digital battlefield, and weapon systems like the new Spike Firefly loitering munition and fifth-generation anti-tank guided missiles.

The systems developed here mesh well with Israel’s multi-year plan, called Momentum, to make the Israeli Defense Forces into a faster, more efficient and lethal army. Israel is relying more on its best technology, such as multi-layered air defense, and its two squadrons of F-35s. It has new units, such as an air force special forces unit and it is doing more combined arms operations combining units and capabilities. It is focusing on what it calls “third circle” threats, such as Iran, as well as how to confront Hamas and Hezbollah. Gozali points out that the future battlefield comes down to being more effective and making every second count. Artificial intelligence can help close that loop.

Read More

Seth J. Frantzman