In 1953, five years after the establishment of the State of Israel, its first prime minister, David Ben-Gurion, wrote what is known as the “Eighteen Point Document,” which has come to be Israel’s only official national security document. In it, he analyzed Israel’s strategic predicament and demonstrated that the geographic, demographic, and military disparity Israel faced with its collective Arab adversaries was staggering; based on the sheer quantitative inferiority, its future looked bleak. His conclusion was simple: we will always be outnumbered, and if we want to survive, let alone win, we would need qualitative superiority – better decision-makers and military personnel as well as better, more technologically advanced weapons of war. This strategy necessarily placed great emphasis on scientific and technological development. Hence, the establishment of the IDF Science Corps – a unit tasked solely with developing weaponry – and the early and heavy investment of Israel’s meager resources into defense technology.
While Ben-Gurion correctly understood that such a strategy would work to Israel’s geostrategic advantage and tilt the balance on the battlefield in its favor, he could not have known, nor did he predict, the windfall that investment in technology would have: a monumental downstream effect on Israel’s economy via the private sector. A recent example is the astounding success of Israel’s cybersecurity industry: as of 2021, 25% of all cybersecurity unicorns were Israeli, 19% of public companies valued at over $1 billion were founded by Israelis, and 10% of the world’s cybersecurity companies were Israeli.
To understand how Israel achieved this level of economic success, we must understand why the IDF started to focus on cybersecurity. The computer and internet revolution in the 1990s brought significant changes to the intelligence landscape. Previously reliant on analog broadcasts, intelligence agencies had to adapt to intercept and decipher digital communications using binary computer code. Traditional methods like antennae for radio and cellular devices became less effective as communications shifted. This shift required IDF Military Intelligence to adapt to the changing landscape and ensure the continued transmission of high-quality, real-time intelligence. By this time, the strategic threat that Iran posed was clear and present, but with the significant geographic distance involved, tapping Iran’s land-based phone signals for contemporaneous intelligence was not an option. However, the rise of the cyber ecosystem presented an opportunity: if Israel could access its enemies’ computers, it could get its hands on emails and any other information the enemy was sending and receiving over the internet. The cyber ecosystem was clearly the best way to collect information anywhere across the globe.
Enter the IDF’s Units 8200 and 81 – both established to intercept and decrypt enemy communications – which began investing intensively in cybersecurity capabilities to gather intelligence. In order to maximize their utility, these units have developed an effective operational lifecycle: the most talented young conscripts in the country are scouted and enlisted, forming the veritable elite from an enormous reservoir of options. While mandatory IDF service usually lasts three years, because these units are so attractive and enjoy such a strong reputation, they are able to make admission conditional on recruits agreeing to sign on for a total of seven years. Nevertheless, because this is far short of the decades of service typical for other intelligence institutions like the United States’ NSA and Great Britain’s GCHQ, Israel’s units have developed an intense, streamlined, and efficient training process. In their seven years, Israeli intelligence soldiers receive experience they cannot gain anywhere else, after which they are released to the private sector to utilize the skills they acquired during their service. Thus, even though the military loses the benefit of these soldiers, the state’s cybersecurity industry gains immensely: this cycle offers a constant stream of highly experienced cybersecurity talent to Israeli private sector to build pioneering and hugely profitable startups.
Indeed, over the years Israel has developed a rich cybersecurity environment that allows private companies to thrive, but the military intelligence tech units remain the main factor in the ongoing success and growth of the private cybersecurity industry.
The Gaza War as a Tech Driver
As with prior defense challenges, one byproduct of the current Gaza War is the opportunity for technological development and even economic gain. On the most basic level, the campaign has provided testing grounds for the various weapons systems developed by Israel’s defense industry. Suffice it to say that the eyes of many foreign governments’ military procurement departments are watching intently. And they are apparently impressed.
A Calcalist article from early November detailed how the successful performance of IAI’s Arrow 3 missile defense system in the current campaign is expected to increase the demand for and accelerate the sale of Israeli missile defense systems around the world.[i] The Houthi attacks from Yemen across the Red Sea provided the backdrop for Arrow 3’s first practical test, and on November 9th it successfully intercepted a surface-to-surface Houthi-launched missile heading for Eilat. During the previous week, the Arrow 2 missile intercepted a ballistic missile launched from Yemen, as well as a long-range rocket that Hamas had launched from Gaza towards the eastern Negev Desert.
Defense industry officials anticipate that the continued operational successes will greatly increase their demand both during and after the current campaign. Thus, while Israel’s enemies seek to attack and undermine Israel’s security, they inadvertently contribute to its economic growth and defense development. This unintended consequence significantly impacts Israel’s military deterrence strategy and sets off a feedback loop of sorts: each major military confrontation not only depletes the enemy’s resources but also elevates Israel’s military technology. This advancement enables Israel to market its technology, thereby gaining both diplomatic and financial advantages. These benefits then feed back into further technological development, propelled by necessity and enhanced by an influx of capital.
An additional area of economic growth will very likely be Israel’s tech industry, and more specifically – artificial intelligence (AI). The Gaza War happened to coincide with an inflection point in AI, and while the IDF had already begun investing in AI, the very specific challenges of the current war suggests that the IDF will double down and focus intense investment on the development of AI capabilities. This will inevitably open the proverbial floodgates for the establishment of start-ups built around this pivotal technological breakthrough.
AI in the Context of the Current Campaign
AI’s role is especially relevant to the most prominent question asked in the wake of the October 7th massacre: “How did we miss this? How did our intelligence apparatus fail to alert us to an assault of such scale?” Israelis take much pride in IDF intelligence capabilities, and it was unfathomable that an attack of this nature and scope could have been missed. Needless to say, there is widespread consensus that our intelligence failed from top to bottom.
However, too many draw the wrong conclusion, suggesting that Israel should divert budgets from the intelligence units to infantry in order to prevent the next attack. While it is obviously important to enhance infantry, it would be a terrible mistake to weaken our intelligence infrastructure. Indeed, it can confidently be asserted that prior to October 7th our intelligence prevented hundreds of terror attacks and that intelligence clearly plays a crucial role in the current war by constantly replenishing the IDF’s target bank.
The right solution to the problem revealed by the October 7th attack is increased investment in intelligence capabilities, with the aim of developing new technologies that can aid in identifying these types of threats faster and provide real-time alerts that enable us to thwart impending attacks. AI can provide Israel with exactly that.
Intelligence derived from the interception of phone calls is a good example. Monitoring and identifying phone calls with high intelligence value introduces real challenges. There is a copious number of phone calls being made by the enemy, most of which are irrelevant, yet all of which still require processing in case one contains critical information. Filtering the phone call that contains the relevant piece of intelligence is like finding a “needle in a haystack.” In addition, not all calls are intercepted perfectly, and some are corrupted or unintelligible. Deciphering them requires special expertise in the Palestinian-Arab dialect, and a solid familiarity with the caller so the information and meaning can be extracted based on the context. This expertise is not something that every intelligence analyst possesses, meaning that there can be a call backlog of many hours to be dealt with by a limited number of experts. The result is clear: slower processing and detection of intelligence. This is a tremendous flaw, because in the intelligence community, while getting the right intelligence item is crucial, it is nevertheless true that “if it’s not on-time, it’s not intelligence.” A priceless alert is meaningless if it arrives after the response window has closed. In the run-up to an attack, where every minute is important, getting critical intelligence items can make all the difference.
In the last seven years, we have witnessed a tremendous leap in AI technologies that can reduce the “time-to-intelligence” metric. The first technology is related to speech-to-text. The development of deep learning algorithms based on neural networks was leveraged for improved speech-to-text capabilities. One notable algorithm is Whisper, introduced in 2022 by OpenAI, which was able to interpret language based on the context, and not purely on word recognition. Today the leading speech-to-text algorithms introduce 85% accuracy. While not perfect, they can transcribe the majority of calls intercepted by intelligence units and compress the time required to process the information – saving precious minutes and even hours.
One additional major advance is the introduction of generative AI, the new generation of Natural Language Processing (NLP) algorithms. The most famous being GPT which powers ChatGPT by OpenAI . ChatGPT is defined as a “foundational model”, meaning a flexible model that can receive an input of various types of text, and then answer questions with information that may be found in that same text. ChatGPT has created much fanfare as it introduced a level of text comprehension previously attributed only to human beings. For our purposes, this type of algorithm can be used to take a phone call that has been transcribed or an email that has been intercepted and provide an automatic answer to a question like “Is there a terror attack planned any time soon?” Essentially, combining speech-to-text algorithms with the latest NLP algorithms can automate the process of transforming data into intelligence. While not perfect yet, and still requiring intelligence expert oversight, it will improve over time and already remove much of the burden.
There is an additional, no-less-crucial realm for AI that has been revealed in the current campaign. Among those who were brutally murdered on October 7th, hundreds of Israelis were considered missing as their status was undetermined. That challenge led to the formation of civilian initiatives that sought to map out the location and status of every missing person. They started collecting pictures of all the missing people and simultaneously began collecting images and videos from the massacre area published on social media by Hamas and the victims. Facial recognition technologies were used to identify the missing persons in the media files and provide an assessment of their status. In the last few years, facial recognition technology has reached new heights, and now boasts over 90% accuracy. These civilian initiatives were only successful because of the latest enhancements in image processing and facial recognition algorithms.
Image processing can also be leveraged for other intelligence uses. Identifying enemy targets for airstrikes is a crucial need for any war conducted by Israel. As Israel abides by International Humanitarian Law, it is only permitted to target unpopulated areas or specific sites known to contain enemy soldiers or infrastructure. In a full-scale attack, like the current Gaza War, known targets are rapidly exhausted and there is a constant need to replenish the target bank. A principal source is aerial or satellite imagery from enemy territory. Image processing can potentially be leveraged to map out enemy infrastructure faster than ever, thereby accelerating its destruction.
It behooves Israel’s leaders to resist public pressure to reallocate budgetary resources from intelligence to infantry. While our intelligence community failed on October 7th, its successes are often unknown and are too many to list. With the right adjustments, it can apply lessons that will strengthen it. Clearly, the foremost adaptation required is in the intelligence “culture,” as it is evident that indications of an impending attack were not taken seriously enough and that there was a general lack of understanding of the enemy’s psychology. Notwithstanding this necessary adjustment, however, technology can also play a major role in rehabilitating the intelligence apparatus.
It is fair to assume that the IDF has already started implementing speech-to-text and NLP technologies to enhance intelligence production, as well as image processing technology for target-mapping. However, the Gaza War has highlighted the importance of intensifying investment in these technologies, suggesting that the defense sector will be spearheading the advancement of AI technology for the foreseeable future. As for the civilian AI industry, Israel is already generating impressive results. Reportedly, there are 850 AI companies in Israel, which place it sixth in the world in total number of companies. Per-capita, it is in second place with 91 companies per million people, just behind the UK. However, from a qualitative perspective, Israel has a low influence on the AI industry as compared to the cybersecurity industry. While companies like AI21 and Mobileye have had a significant impact, Israel still has not been able to position itself as a world class powerhouse for AI technologies. If, as predicted, the IDF invests more heavily in AI technologies, a stream of highly talented and experienced AI professionals is bound to emerge from the IDF and boost the civilian AI industry. This is precisely what Israel needs to become – the groundbreaking leader in AI that it already is in cybersecurity.
This war, despite all its horrors, presents Israel with an immense opportunity. Soldiers serving in units 8200 and 81 are applying their skills in a high-pressure environment using the most cutting-edge technologies. Reservists, together with their colleagues from the Navy, Air Force, and infantry, will eventually return to their civilian jobs in high-tech sectors. They will bring with them invaluable insights about what worked and what could have been more effective. In addition to having directly experienced the harsh realities of war and demands of survival, these Israelis are equipped with entrepreneurial drive, scientific expertise, and access to capital. These factors, combined with their patriotic mindset and a sense of urgency, position them to become the next generation of outstanding innovators. Their contributions have the potential to catapult Israel to a leading position in AI technology, which would not only enhance the nation’s security but also ensure its continual prosperity.
Alon, CEO of PointFive, co-founded IntSights, leading to a $350M sale to Rapid7. He wrote “The Battle for Your Computer” and founded “Kanfey Kesef” and “Step2Tech” for social impact. He’s also an active angel investor and startup advisor.
[i] https://www.calcalist.co.il/local_news/article/sknzj3ix6.
