Top Israeli intel officer goes where no one’s gone before. And you can find it on Amazon

Brig. Gen. Y., commander of Unit 8200 in the Israel Defense Forces’ Military Intelligence, has apparently set a precedent. It’s difficult to recall another case in recent decades of a senior officer in active service publishing a book dealing with concepts and working methods in his professional realm. And when the book deals with one of the hottest topics in the field of technology – the use of artificial intelligence – and is even for sale on Amazon, it’s definitely a special case.

The rather slim volume, published only in English, is entitled “The Human Machine Team.” It’s based on a research study that Y. wrote before taking up his current post, when he was on academic leave at the National Defense University in Washington. The blurb describes Y. as “one of the world’s leading managers in the field of Artificial Intelligence,” a recipient of “the prestigious Israel Defense Prize,” an “expert analyst” and the “commander of an elite intelligence unit.” The blurb states that the book “unveils the secrets to creating synergy… that will revolutionize our world.” In fact, the book discusses complex subjects for readers who are not necessarily technology or intelligence experts, without being able to enter into the in-depth details of the nature of intelligence work. But Y. appears to have overcome these hurdles safely.

At present, Y. writes, “we are only at the threshold of the acceleration of the Digital Era.” The result is that large companies and organizations, including intelligence agencies, will be compelled to enhance the joint work of researchers and computer systems. “A machine can use big data to generate information better than humans. However, a machine can’t understand context, doesn’t have feelings or ethics, and can’t think ‘out of the box.’ Therefore, rather than prioritizing between humans and machines, this book is about The Human-Machine Team (‘super-cognition’) and about the collaboration between human intelligence and artificial intelligence.”

The book promises to provide a new perspective on how “to lead our nations and organizations to a future that has already become the present.” Interaction between human and machine can resolve “national security threats and challenges, lead to victory in war and serve as a growth engine for humankind.” These are ambitious goals, and the book is not a manual for achieving them, but it’s still intriguing, not least because the author is a person who’s involved with the country’s most sensitive systems.

Y. begins the book by describing the October 2018 attack on a Pittsburgh synagogue. An antisemitic terrorist, Robert Bowers, shot and murdered 11 worshippers and wounded six others. An hour before the attack, he wrote a post in social media in which he accused a Jewish organization that aids foreign refugees of “bring invaders in that kill our people” and warned that “I can’t sit by and watch my people get slaughtered.”

Y. continues: “Now imagine that an hour before the murderous attack, a red light flashed in the synagogue with a sign that read ‘Beware: there is an unfamiliar cellphone in the synagogue compound.’” Let’s imagine further, the author goes on, that “a camera – equipped with facial recognition – identified Bowers, alerted police to an imminent threat and activated an alarm inside the synagogue.” A few months before the assault, the murderer was issued a license to carry firearms, even though he had previously published sharp attacks against Jews. Imagine further, then, that his post had been monitored in time and crossmatched with the available information. (In actuality, Bowers was not known to the authorities and was not on a watch list.)

According to Y., the potential to identify him in advance could have existed if a “human-machine team” had dealt with the situation. Y. notes that “not even 20,000 analysts with 20,000,000 years of time” could “connect the dots and build the profile of a suspect like Bowers.” However, it can be done by means of a computerized data scan. Our era, he writes, is characterized by growing numbers of attacks by “lone wolves” who draw inspiration from inciteful material on social media networks. But there are more effective ways to deal with them.

The solution proposed by Y. has already been is use for some years, of course, but not in Western democracies. Israel makes use of similar methods in the territories, following lessons gleaned from the “stabbing intifada” in the fall of 2015.  

Y. recalls how funerals followed one another in rapid succession during that period. He and others were constantly telling themselves that there must be some sort of signal which, if detected, could prevent an attack. The sense of a crisis grew more acute as time passed and intelligence was of little help. Finally, intelligence personnel realized that the way they had done their job for decades was insufficient to meet the challenge. The solution was found in teamwork. The human role was to provide examples as to the character of previous attacks; the machine role was to scan the databases. By that method, dozens of attacks a month were prevented.

Machines learn from experience

The system’s components are built from the compilation of large databases together with an increasingly rapid ability to detect anomalies in the data (such as locating the cellphone of a person with a dangerous background in a sensitive place for an attack), on which deterrence can be based. Y. describes the progress that had been made in the ability of computers to establish human-like awareness and to think. He does not foresee a dystopian future in which machines will seize control of humans and make strategic decisions. However, he writes about the “development of cognitive ability,” such as the collection and analysis of information, its processing and reorganization, and the ability of machines to learn from experience in order to give feedback, make evaluations and even arrive at conclusions.

Y. describes “synergetic learning between humans and machines,” which will make it possible to understand and analyze subjects that the human brain cannot process on its own. He mentions a term from his youth as a yeshiva student – “havruta” – which refers to learning together and drawing on knowledge that is acquired from an encounter between different points of view. 

The point of departure for a discussion of these processes, he writes, is that the past four or five decades were only “‘foreplay’ for the big changes that are about to occur in the coming years.” Thus, the machine’s “ability to perform human cognition,” combined with the “ability of the Human-Machine Team to learn together and think together will create a new world.” 

The examples he cites are from the world of national security but are relevant, the author argues, to other fields, such as the economy, health care and personal security. In short, “AI is changing the basic rules.”

He also discusses the growing difficulty of protecting borders from terrorist attacks and the specific Israeli problem of dealing with Hezbollah’s arsenal of missiles and rockets, which are hidden among a civilian population. Uncovering the rockets is difficult, Y. writes, and it’s highly problematic to attack the launchers without inflicting casualties on a large number of innocent civilians. He notes that in the “Lebanon War in 2006, more than 40,000 rockets were fired by Hezbollah into Israel over a period of 33 days. For the next war … Hezbollah plans to fire thousands of rockets against Israel every day,” some of them high-precision projectiles. 

AI will be mustered for the location and identification effort. In the meantime, some of these processes were already manifested in Operation Guardian of the Walls in the Gaza Strip last May. In the decades ahead, Y. writes, analysts who analyze language (recordings and computer printouts) and aerial photographs will gradually disappear and be replaced by machines. Already within five years, there will be no further need for 80 percent of the listeners in intelligence organizations, because that job will be handled by computers. A large part of intelligence gather will be conducted by thousands of drones.

The military challenge facing Israel will be to forge a tighter connection between the “fire” effort (the attacks) and the intelligence effort (the data), in a manner that will locate targets for an attack in real time, while causing minimum damage to civilians. The pressure on the enemy in a war will make them want to end it earlier, but “there is a human bottleneck,” because humans do not have the ability to process so much data. The answer lies in the human-machine team.

Like Chief of Staff Aviv Kochavi, Y. describes a “targets machine,” based on intelligence, as a “game changer” that could obtain victory for Israel in a war. The AI revolution, he writes, gives humans new glasses to look through for every challenge and mission.