You are here

Israel's national AI plan unveiled

Prof. Isaac Ben-Israel and Prof. Eviatar Matania head an effort to make Israel an AI power by replicating the success of government investment in cybersecurity. Making Israel one of the world's five leading countries in artificial intelligence (AI) will require a NIS 1-2 billion annual investment by the government, out of the NIS 10 billion invested in this sector in Israel, according to the recommendations of a committee headed by Maj. Gen. (res.) Prof. Isaac Ben-Israel and Prof. Eviatar Matania.

Parts of the committee's draft report were revealed to "Globes", and were presented in brief as part of the AI Week held at Tel Aviv University in cooperation with the Israel Innovation Authority and Intel. The committee has 15 subcommittees, on which 300 senior people from the government, the IDF, institutions of higher education, and the technology industry serve. Its recommendations are slated for submission to the government in January 2020, if there is a government at that time.

AI is a key technology in both the defense and the economic spheres. Dozens of countries have begun devising a national strategy in this area. China, for example, has a $141 billion plan for developing its core AI industry (as opposed to overlapping and related industries) up to 2030. Russian President Vladimir Putin has said, "A nation that leads in AI will control the world."

The world's expectations from AI technology and its applications are sky high. Consultancy McKinsey & Co. predicted in April 2018 that the potential business value that can be produced from AI technology is $3.5-5.8 trillion annually. "Business value" is a term used by research companies to express the full value created by the use of technology, including savings in work hours, for example.

Most of the technological revolutions now taking place or about to take place are based on AI. AI technology uses algorithms to analyze large quantities of data, detect recurring patterns, derive insights from them, and adjust themselves to changes and operate accordingly. For example, Mobileye, acquired by Intel for $15.3 billion, is actually an AI company, because it developed algorithms enabling an autonomous vehicle to cope with complex situations. Such algorithms are trained using data gathered from real situations, and develop and improve as more data are accumulated.

AI technologies make it possible to use the accumulated mountains of data and turn them into practical knowledge that can also contribute to business efficiency, for example.

Learning from the cyber program's success

Ben-Israel and Matania, who established the committee in Israel as a government initiative, with Netanyahu's support, were behind a similar initiative early in the decade for promoting cybersecurity technologies. This initiative is credited with making Israel a cybersecurity power.

The current plan's goal is to apply the formula that proved successful in cybersecurity to AI. In contrast to the cybersecurity sector, however, Israel did not become aware of AI early enough, and entering the race with redoubled force is now a matter of some urgency.

The plan's expected cost is based on the idea that in order to propel Israel to the forefront in the sector and overtake countries that have already been active in it, a large initial government investment amounting to 10-20% of total annual government spending on civilian research and development is needed.

Research and development spending in Israel is equal to 4-4.5% of GDP, with 20% of this spending going for technological R&D. Assuming that total spending on technological R&D is NIS 60 billion a year, including NIS 10 billion from the government, the required government investment for AI under the plan will be NIS 1-2 billion. For the sake of comparison, NIS 2.5 billion was invested in the entire cybersecurity program.

Based on the experience with cybersecurity, Ben-Israel expects the investment to pay for itself in the Israeli economy. Investment in business R&D in the cybersecurity sector in Israel is now 18% of global investment, while Israeli exports already account for 10% of global cybersecurity exports. Even if Israel's share of the global AI market is only 1%, the return on the investment will be huge.

Founding a special authority and computer infrastructure

The committees' recommendations involve many aspects of business and government activity, in the realization that there will be no area that will not be deeply affected and disrupted by AI technologies in the future. One of the spheres that the committees recommended as a focus is digitization of government services, both between different government ministries and between the ministries and the people that they serve.

The committees discovered that AI could solve some of the problems that cause accessible government projects to fail. The committees recommended national projects that Israel should undertake, one of which is shortening waiting lines in the health system.

In order to implement the plan, the committee recommended establishment of a coordination agency that would be part of a government ministry, similar to the National Cyber Directorate, which is currently part of the Prime Minister's Office. The Cyber Directorate is the result of Ben-Israel and Matania's national cybersecurity plan.

A coordinating agency is needed because of the large number of authorities involved in the local technology industry: the Innovation Authority; the Ministry of Communications; the Ministry of Science, Technology, and Space; Digital Israel (part of the Ministry of Labor, Social Affairs, and Social Services); the Council for Higher Education in Israel's Planning and Budgeting Committee; and of course defense agencies, headed by the Ministry of Defense Administration for the Development of Weapons and Technological Infrastructure.

In order to speed up research and development, Israel will have to build suitable computer infrastructure within its borders. The committees recommend the publication of a tender for building an AI cloud, with likely involvement of companies like Amazon, Google, Nvidia, and Intel. Senior sources in the industry predicted a few months ago that spending on the computer infrastructure alone, without storage and communications, would be NIS 500 million for five years.

The smart cities sector also received special attention, with a recommendation to classify an Israel city, probably Tel Aviv, as a "trial city" for smart transportation and autonomous vehicles. The committees recommend a special emphasis on metropolitan traffic control centers and smart traffic lights.

The committees recommend the founding of 4-6 relevant research centers in universities that will increase the number of researchers in the sector, from first degree students to leading researchers. The committees are proposing a "professor model" in order to cope with giant companies snatching up experts. Today, most intellectual property is in the hands of researchers, and the committees recommend a change in this state of affairs. Other recommendations are adding a matriculation exam in AI and a compulsory course in research methods using AI.

At the same time, the committees recommend increasing the number of foreign research students in order to substantially boost the number of publications in the field (Israel was only in 18th place worldwide in per capita publications in 2017). Another recommendation is a reform that will enable senior researchers to substantially increase the amount of time they are entitled to devote to work in industry, rather than at the university. This measure is designed to solve a problem: neither academic institutions nor local industry are capable of competing with the huge salaries offered to experts by the technology giants.

The most promising area in which the committees recommend that AI should be cultivated is agriculture. According to the committees' draft report, Israel can potentially become a global leader in the field, because of the know-how it has already accumulated. Israeli know-how can assist in optimal use of natural resources and minimizing environmental damage. It is argued that Israel will be better able to cope with the environmental challenges facing it, while at the same time contributing to the entire world.

The focus on agriculture in the committees' recommendations stand out, given the regulatory restrictions and difficulties involved in medical research resulting from privacy issues, among other things. Agriculture is less restricted in these aspects, and can therefore be a fruitful field for research, particularly in plant diseases.

The point of departure is that as more AI capabilities are attained, our dependence on computers will increase. The systems to be developed must therefore be secured very well. Ben-Israel wants to call the venture "secured intelligent systems."

Finally, the committees did not ignore the question of ethics. The committees' recommendations do not provide researchers with the ethical criteria themselves, but the recommendations provided give them a "ruler" they can use to make sure that they have gone through all of the appropriate processes for weighing ethical considerations.

Private investment flowing into the sector

Israel is today a global leader in the number of AI startups, with nearly 400 companies, third in the world after the US and China, according to figures from the Asgard Capital venture capital fund. According to figures from Start-Up Nation Central, the number of Israeli concerns, including research centers, developing AI technologies or making significant use of them was estimated at the end of last year at 1,150. Investment in companies in the field last year totaled $2.25 billion, 37% of investment in the technology sector, even though the proportion of companies in the field was 17%.

International companies also conduct substantial AI activity in Israel. For example, Intel operates a development center for AI technologies in Israel for its own internal needs. The center is headed by Itay Yogev, who founded the unit 10 years ago. It now has nearly 200 employees most of them in Israel. Another example is the Israeli startup, which uses AI to streamline work by employees using the organizational software at their places of work.

Intel also has extensive AI activity in hardware. It is developing AI accelerators from Israel that are being integrated in the core of an ordinary chip. Intel is developing AI processors for data centers used by Facebook, and of course also has Mobileye developing AI technologies for an autonomous vehicle.

There are also several startups in Israel, such as Hailo Technologies and NeuroBlade, developing AI chips in competition with the global giants, and many other startups in the sector. In the first quarter of 2019 alone, startups in the sector raised $600 million in 51 deals, accounting for over a third of all the capital raised in the quarter.

Although the investment figures are impressive, the private sector's investments in startups and mergers and acquisitions are not the most important thing. The committees' goal is to determine a way for the education system and academic institutions to supply the required number of engineers, so that the government, the IDF, and industry can obtain employees of sufficient number and quality, not only in order to succeed, but to prosper in the long term.

Uri Berkovitz