Israel developing AI 'super-brain' to run automated attacks in battlefield

A mock demonstration of the tool shows a tank with sensors and radars around it
A mock demonstration of the tool shows a tank with sensors and radars around it

Israel is developing an AI "super-brain" that uses an array of high-powered sensors to help tanks and robots patrol battlefields and find enemy targets, as it looks to a new age of warfare.

The artificial intelligence, dubbed Athena after the Greek goddess of war, is in an early stage of development at Israel Aerospace Industries (IAI) but could be deployed within the next decade.

The Daily Telegraph was given exclusive access to the Athena programme, including a mock demonstration of the “brain” being used by tanks in a combat scenario.

In the scenario, Israeli Carmel tanks were fitted with the smartphone-sized AI, which collected data from infrared and radar sensors, among others, to tag enemy fighters hidden underground and in buildings on the battlefield. In an instant, the data was beamed to the commanding officer and turned into a “battle menu” with the best methods of attacking the targets.

The Athena can also be plugged into the tanks’ fire control and manoeuvre systems, which would allow it to attack targets automatically, though the commanding officer makes the final decision on how to proceed.

The Telegraph understands that IAI, Israel's state-owned defence firm, is also looking at fitting the Athena onto robotic vehicles that could automatically patrol border fences in search of intruders.

Another image from the mock-up shows armoured vehicles patrolling
Another image from the mock-up shows armoured vehicles patrolling

The country hopes AI defence systems will make its armies more efficient, as the machines can analyse a battlefield and generate a tactical battle plan much faster than a human mind. Senior officers in the Israel Defence Forces (IDF), which is also exploring AI warfare, said some aspects of the technology could be ready for real-world use as early as next year.

“We are not very far from the basic level of getting an automatic awareness picture [of the battlefield],” said Colonel Eli Birenbaum, the head of the IDF’s military architecture department. “If you ask me [when it will happen], at the end of 2021, I think with 90 per cent accuracy that the answer is positive.”

Col Birenbaum said one of the key challenges in the future will be persuading soldiers to “trust the machine,” drawing comparisons with safety concerns over self-driving cars.

The Athena device is just one of a number of “AI weapons” in development around the world, though the projects are generally shrouded in secrecy. They include drones and missile targeting devices, while experts point out that for many decades, armies have used weapons that automatically attack people, such as the landmine.

A computer image indicates where the AI system would fit on a tank
A computer image indicates where the AI system would fit on a tank

“There is a bit of an arms race underway between the US and China, while the Russians are also doing a lot of experiments but are quite a bit further behind,” explained Dr Jack Watling, an expert on land warfare at the Rusi defence think-tank in London.

“It is realistic, and very much available, but not necessarily deployable. And this is partly because militaries are not necessarily sure of the ramifications of using it,” he added.

Israel is not the only Western ally exploring AI warfare. In the United States, military cadets are programming tanks with algorithms to pop balloons representing enemy soldiers.

Meanwhile, last November, the head of the British armed forces speculated that “robot” soldiers could make up around a quarter of the military by the 2030s. General Sir Nick Carter made the comment in an interview about UK defence funding, though he said he was not setting any particular target on a robot quota.

It has been claimed that Israel deployed an AI-powered machine gun late last year to assassinate one of Iran’s top nuclear scientists. However, it has not responded to claims it was behind the attack and typically declines to comment on overseas military action. The scientist, Mohsen Fakhrizadeh, said to be the head of Iran’s nuclear programme, was reportedly killed by a satellite-controlled machine gun, which used artificial intelligence to target him.

Over the weekend, tensions rose between Israel and the Palestinian territories after more than 30 rockets were fired from the Gaza strip into Israeli territory by Islamist militant groups.

The rockets did not cause any deaths or injuries, though they did damage some property. The rocket barrage appeared to be in retaliation for clashes between Palestinians and an extremist Jewish group in east Jerusalem on Friday, which left more than a hundred people injured.