AI is at the center of the current industrial revolution, and its significance is likely to increase in the coming decade. Many people believe that AI will lead society to a new technological era, propelling a revolution even more momentous than previous industrial revolutions. In recent years, the ability of computer systems to use statistical algorithms to simulate human cognitive processes, such as learning and self-correction, has grown leaps and bounds. These systems are also capable of harnessing enormous amounts of information accumulated in various databases around the world in order to gain valuable insights.

In recent years, thanks to a surge in AI technologies and their enormous economic potential, many countries have devoted a great deal of effort and resources to the race for technological leadership in AI.

Source: Politics = AI, Tim Dutton, 2018 + European Commission data 2019

The 2018-2019 Innovation Report¹Israel Innovation Authority, “The Race for Technological Leadership: The Advent of the Smart Machine Era.” Israel Innovation Authority Report, 2018-2019 included a survey conducted by the Israel Innovation Authority that delineated the heavy investment in AI made by nations such as China and other East Asian countries. Since then, the US, Europe, and other countries have revealed strategic plans aimed at advancing AI. Figure 5.1 shows a timeline of countries that have thrown their hat in the AI ring and announced national plans dedicated to AI.

A small country like Israel cannot realistically compete with the enormous investments made by countries like China or leading global high-tech companies like Amazon or Google. The Israeli innovation ecosystem has, nonetheless, many relative advantages that allow Israeli companies to assume technological leadership, such as entrepreneurial spirit, audacity, and flexibility. Another advantage that Israel possesses, exclusive to AI, is the large amount of raw digital information in national databases, particularly in healthcare. Access to large databases is critical for young companies developing machine learning processes or fine-tuning algorithms, and is essential to the technological development of AI. Such access could contribute to Israel’s capability to lead this field.

Israeli Leadership in AI

The growth of AI technologies is reflected in investments made in the field. In 2011-2019, investments in Israeli high-tech AI projects increased by a factor of 12.5, from 305 million dollars to 4 billion dollars. Figure 5.2 shows that in 2019, 42% of the total sum invested in Israeli high-tech went towards AI technologies.²IVC data processed by the Israel Innovation Authority

Source: IVC data processed by the Israel Innovation Authority

In recent years, Israeli high-tech has heldMay:leading global position in the field of AI. Israel is now among the world’s top three countries operating in the field of AI, following China and the US.³The Global Artificial Intelligence Landscape, Asgard Moreover, according to research company CB Insights, Israel ranks second, after the US, in its number of leading AI start-ups.⁴CB-Insights top startups ranking It also ranks high in the number of companies that develop infrastructure technologies for AI such as special- purpose chips, infrastructure algorithms, and complex systems for the acceleration of computing.⁵MMC Ventures & Barclays UK Venture, “The State of AI 2019: Divergence”, 2019

Source: IVC data processed by the Israel Innovation Authority

Giant tech corporations, such as Google, Amazon, Apple, Netflix, Facebook and Microsoft, hold a dominant global position in AI research and in the development of core technologies that employ AI.⁶OECD and EU-JRC, World Corporate Top R&D Investors: Shaping the Future of Technology and of AI, 2019 This has ramifications on Israeli companies operating in the field. On the one hand, large multinational corporations leave little room for small competitors trying to develop core technologies for AI, as this requires tremendous expertise, abundant resources, and proximity to academic research.⁷OECD and EU-JRC, World Corporate Top R&D Investors: Shaping the Future of Technology and of AI 2019 On the other hand, the dominance of large corporations also offers the Israeli innovation ecosystem a significant advantage and contributes to its reinforcement and enrichment in several ways. The massive presence of roughly 90 R&D centers of multinational corporations working on AI in Israel (out of around 400 foreign R&D centers operating in Israel)⁸Startuphub.ai 2019 positions Israel at the forefront of global technology. The large international presence produces a flow of information between the multinational corporations and local start-ups, and exposes Israeli companies to the cutting edge of the field. Multinational corporations also contribute to the development of local human capital, by giving local engineers and leaders technological and managerial experience, which enriches Israeli high-tech.⁹For more information, see: Multinational Corporations’ R&D Centers in the 2017 Innovation Report, the IIA

Maintaining Israel’s Lead in AI

Despite the rapid growth of Israeli high-tech, and of AI in particular, strategic analysis performed by the Israel Innovation Authority has identified four obstacles that can slow future growth in the field:

• A shortage of human capital skilled in the field of AI
• Limited access to public and government databases for use by new companies
• Inadequate supercomputing infrastructure for the development of advanced technologies
• Slow-changing regulation and a lack of ethical guidelines

Insufficient government support on these issues may result in the deterioration of Israel’s high standing in the development of AI technologies. The Innovation Authority, in collaboration with its partners in the government and the industry, is working to create solutions to these challenges.

A Shortage of Skilled Human Capital in AI

Advances in AI have led to a sharp global surge in demand for skilled professionals at every level of the professional pyramid: analysts and data scientists, machine learning specialists, and algorithm developers. Figure 5.4 mimics a LinkedIn graph showing the demand for data scientists and AI professionals, which increased more than any other profession in the past decade. Demand for AI specialists had the highest growth rate, with an annual increase of 74% in the recruitment of AI professionals in the US.¹⁰LinkedIn, “2020 Emerging Jobs Report”

According to Israel’s 2019 High-Tech Human Capital Report by the SNC and the Israel Innovation Authority, 7% of available tech positions in the industry, some 1,300 positions, are for AI or data sciences professionals.¹¹Israel Innovation Authority and Start-Up Nation Central, “Human Capital Survey Report 2019”, working version

Israel ranks 18th out of the 42 listed countries in its number of AI scientific publications per capita.¹²Ranking of countries that published articles on AI in the academic database Scopus in 2013-2017. Source: The Samuel Neaman Institute for National Policy Research, Artificial Intelligence, Data Science and Smart Robotics. First Report, 2018, p. 13 While the number of Israeli publications on the topic has grown annually, the growth rate is slower than other countries, and the proportion of Israeli AI research is shrinking.¹³Ibid, p. 15. Should this trend continue, Israel’s ability to lead global AI research will be limited. The amount of research is affected, in part, by the high demand for data scientists in the private sector, which makes it difficult to recruit and retain academic faculty in AI related research fields. This, in turn, hinders academic research and instruction, and impedes the ability to create future generations of leading researchers in academia, who are required to develop and transfer knowledge from academia to industry.

Furthermore, Israel is also losing skilled human capital in AI due to migration, illustrated in Figure 5.5.

In recent years, Israeli academia has responded to the increasing demand for skilled AI professionals, and launched the National Program for the Advancement of Data Science in Higher Education accordingly, among other endeavors.¹⁴In 2018, 17 new data science education programs were submitted for approval by the PBC (the Planning and Budgeting Committee). In 2019, the PBC approved the National Program for the Advancement of Data Science with a budget of approximately 150 million shekels, which includes the funding of scholarships for doctoral and postdoctoral students However, more must be done, and the long term policies must be complemented by short-term, targeted training programs.

As will be detailed later in this chapter, the Innovation Authority is working on non-academic training programs to provide alternative career paths in AI, such as coding boot-camps and the Workshop for Advanced Technology Training.¹⁵For more information, see chapter 7 of this report, The Public Social Arena

Limited Access to Data

The development and training of AI algorithms require access to large databases that facilitate insight mining and machine learning. These large databases are usually owned by the government or by multinational corporations like Google and Facebook. The Israeli government has thousands of national databases, but only around 700 of them are accessible to the public. Most are not normalized, with non-validated and unlabeled data, and therefore do not qualify as high-quality data.

Enabling access to high quality databases requires work on several fronts simultaneously: organizing data in existing databases, synchronization between databases, data anonymization,¹⁶Data anonymization is when a database that contains personally identifiable information undergoes a process that includes the extraction of identifying information, ultimately creating a new database that does not allow the exposure of personal information, and whose use does not violate the privacy and confidentiality of the people included in the original database creating synthetic research databases based on existing databases, collecting data and building new databases, establishing regulation, and building centralized computing infrastructure to manage information between all entities who will make use of the data.

Regulation in Israel is not currently suited to keep up with technological developments. Start-ups’ access to databases is often blocked unnecessarily, slowing product development. The Israel Innovation Authority is working in collaboration with a variety of government entities, including Digital Israel and the Ministry of Justice, to improve regulation standards for data. The Authority is also working to advance data unification processes and the standardization of national databases that have applicable value.

Supercomputer Infrastructure

Other forms of infrastructure that are critical for the development and training of AI algorithms are computational capacity and High-Performance Computing (HPC), which is necessary for meeting the demands of data processing. Currently, there is no such advanced infrastructure in Israel, despite its importance to both industry and academia.

Figure 5.6 shows a list of the countries with supercomputing centers – Israel does not appear on the list.¹⁷TOP500, Supercomputer List – November 2019

Source: TOP500 List Supercomputer

The lack of supercomputing infrastructure can also be seen in the Global AI Index. The index ranks Israel third (out of 54 countries) in AI investments, fifth in R&D, but only 36th in infrastructure.¹⁸18 “The Global AI Index”, Tortoise Media

Local supercomputing infrastructure would lower expenditures for Israeli companies and would increase their capacity to advance independent R&D. While giant companies maintain their own infrastructure, small companies must use cloud computing services and overseas storage at a high cost, and are limited to generic, inflexible tools, available in the private sector. In addition to high costs and inflexibility, relying on foreign cloud infrastructure also causes slow response time – the time it takes between running the data and receiving results. Another problem arises when information is sensitive and classified (such as personal medical information or information that is classified for security purposes), or when information cannot or should not be transferred to foreign infrastructure.

The Israel Innovation Authority believes that the lack of local AI supercomputing infrastructure is limiting Israeli high-tech’s competitive advantage and that establishing this type of infrastructure in Israel is vital. This is further corroborated by the fact that many governments worldwide are currently working to provide infrastructure to academia and local industry.

Changes Needed to Privacy and Ethics Policy in AI

Implementation of smart systems raises ethical concerns that call for clear regulatory definitions. Using large databases poses privacy challenges that mandate information classification by sensitivity levels. Moreover, at times, it is not clear how AI systems make decisions. The responsibility of the manufacturer or the user for the machine’s ‘independent’ activity, in the case of critical error, remains unclear (for example, in an autonomous car accident).

Israel Innovation Authority Advances Israeli AI Leadership

The Authority, in collaboration with relevant government entities, employs a variety of tools to promote growth in AI and to remove obstacles facing the field:

Israel Innovation Authority Backs Groundbreaking AI Projects

In order to continue encouraging the successes Israel has already experienced in the field of AI, the Authority provides financial aid and support to companies with groundbreaking AI technologies. In 2019, the Authority provided grants to 219 companies working on AI technologies for a total sum of roughly 415 million shekels. These grants were mostly provided for AI infrastructure projects and in sectors underfunded by private investment.

Developing Skilled Human Capital for AI

In an effort to address the increasing shortage of human resources in AI, the Authority believes that both academic and non-academic training is required, and that foreign specialists should be encouraged to come to Israel. To this end, the Authority is operating various tools to train employees for in-demand AI professions.

The Authority recently launched a new program to support professional training in high-tech: The Workshop for Advanced Technology Training, which focuses especially on AI. The innovative model will allow companies to achieve relatively quick results by propelling existing developers and engineers in the industry to become AI specialists, and will encourage information sharing and the creation of an AI community. Four of these workshops are slated to be held in 2020-2022. Each workshop will train at least 200 employees, with a total government investment of 16 million shekels. The Authority will also continue to fund professional training via coding boot-camps, and will add professional training workshops to R&D consortiums focused on AI, which are funded by the authority through its MAGNET (Generic Technologies R&D Consortiums) programs. Two coding boot-camps in data science, which receive financial backing from the Authority, are already operational.¹⁹For further details, see chapter 7 of this report, The Public Social Arena

Technology Infrastructure for AI – Databases, Software and Computing Power

The Israel Innovation Authority believes that there is a pressing need for a national supercomputing infrastructure for the industry, academia, and security sectors. In order to gauge the scale of infrastructure that would meet the needs of its users, the Authority, together with its partners in the TELEM forum (the National Infrastructure Forum for Research and Development)²⁰The TELEM forum includes, in addition to the Israel Innovation Authority, the Ministry of Science, the Ministry of Finance, the PBC, and the Administration for the Development of Weapons and Technological Infrastructure for the Ministry of Defense referred to potential partners who specialize in this field. These partners include large companies, researchers and professors from all over the country, security and government entities, and specialists from similar companies overseas.

Infrastructure characterization will focus on the use of technologies developed in Israel in order to ensure availability and high-quality engineering support, while reinforcing commitment to the project and bolstering the financial contribution to Israeli high-tech.

In addition, the Authority is backing a range of MAGNETs (collaborations between companies and academic researchers for the development of generic technology) and user organizations in order to facilitate the development of dedicated databases and technological tools that support AI, such as software infrastructure and code libraries. These consortiums currently operate in a variety of fields, such as digital health, transportation, and natural language processing. The total investment in this infrastructure is slated to reach a total of 320 million shekels throughout the duration of the program.

Advancing Supportive Regulation in Collaboration with the World Economic Forum

The Israel Innovation Authority works in several ways to advance regulation that supports innovation in the Israeli high-tech industry. To this end, one of its primary tools is the operation of the center for the regulation of innovative technologies in collaboration with the World Economic Forum (WEF).

In 2019, Israel joined a network of WEF centers that advocate regulatory advances to promote innovation. The Israeli center, which recently began operating under the auspices of the Israel Innovation Authority, will work with local regulators to adjust local regulation to better accommodate future technologies. In concert with the network administrator in the US, the Israeli center has begun exploring regulation that would encourage a competitive, supervised and sophisticated market. Being a part of this network allows Israel to advance regulatory changes while gaining insights from regulatory innovation across the globe. Projects undertaken by this global network include civilian use of drone systems, IoT regulation, improved transportation management capabilities, and auxiliary applications in medical informatics. In 2019, the Israeli center concentrated on assisting regulators in smart transportation and digital health.

At the same time, in September 2019, Israel adopted ethics guidelines for responsible R&D in AI as a member of the OECD. These guidelines ensure the protection of users’ privacy and ready access to public databases.

In conclusion, AI technologies are projected to have an unprecedented impact on the global economy. The Israeli innovation ecosystem, at the forefront of technology and knowledge, is in prime position to lead this field. For Israel to realize its full potential, it must make a concerted effort to meet the challenge of a lack of human capital, improve access to databases, enhance sub-par computing infrastructure, and amend the regulatory environment. The Israel Innovation Authority is working to address these challenges and maintain Israel’s leading position in the global race for technological leadership in AI.