The Italian Ecosystem¶

The Italian AI ecosystem is composed of four categories of actors: 1) research communities 2) knowledge-transfercentres 3) technology and solution providers, and 4) private and public user organisations. This section reviews each category’s relative strengths and weaknesses.

1. Research Communities

Italy has a vibrant AI research community, which contributes to the Italian AI ecosystem in four ways:

• Research: Italian researchers are active in both fundamental and applied research, thus covering all fields of AI. In particular, Italian research is clustered around the following topics: Machine/Deep Learning, Computer Vision, Natural Language Processing (NLP), Data Mining, Big Data analytics, Embedded AI, Human aspects in AI, Knowledge Representation and Symbolic Learning, Decision Support Systems, Agent-based systems and Trustworthy AI. Research in AI is mainly carried out in research labs of universities and public research centres as in the National Research Council (CNR) - and in research foundations such as Fondazione Bruno Kessler and the Italian Institute of Technology. As we shall see later, this community has a widely recognized international standing in terms of quality and quantity of its research output.
• Education and training: Italian universities offer more than 200 AI curricula spread over roughly 50 universities. To further boost this ecosystem, Italy launched in 2021 the National “Artificial Intelligence” PhD Programme (PhD-AI.it), one of the largest and most ambitious PhD programmes in artificial intelligence worldwide. The programme involves over 50 universities, three public research bodies and three research organisations with a view to train researchers, innovators and professionals. The PhD-AI.it is composed of five federated PhD courses, which share a common core curriculum in AI foundations and developments. Each course has an area of specialisation related to a strategic sector for AI application: health and life sciences, agrifood and environment, security and cybersec, industry 4.0, and society. To date, the programme has issued 200 PhD scholarships with a budget of €16 million [1] .
• Assets: the Italian ecosystem hosts several top-tier research infrastructures, in terms of computational power and data lakes. Examples are the CINECA-INFN Infrastructure for HPC, CNR -High Performance Artificial Intelligence Center HP-AI, the IIT HPC infrastructures devoted to the new 2020-2023 Strategic asset in AI and Machine Learning. These are all capable of providing computing facilities and AI competences that allow scientists of any discipline to realise complex and computationally demanding AI experiments.
• Communities: Italian researchers participate in all major AI international research networks, including the top rated EU networks, such as CLAIRE [2] , ELLIS - confederating several AI laboratories all over Europe [3] - and the networks of excellence under the EU Horizon 2020 ICT-48 portfolio, namely HumanE-AI-Net, TAILOR, AI4MEDIA, ELISE and VISION. Italy is one of the founding members of Global Partnership on AI (GPAI) as the result of an idea developed within the G7, under the Canadian and French presidencies. This partnership aims to bridge the gap between theory and practice in AI. Italy has also a remarkable tradition of national scientific research associations, such as AIxIA, CVPL and AILC. In particular, AIxIA is the Italian Association of Artificial Intelligence, founded in 1988, and a member association of EurAI, the European Association of Artificial Intelligence. Many Italian researchers have covered important roles in the EurAI board. CVPL is the association of Computer Vision, Pattern Recognition and Machine Learning founded in 1983, devoted to theory and application of AI for multimedia data, connected with the International Association of Pattern Recognition. AILC is the Italian Association of Computational Linguistics, mainly devoted to research in natural language processing.

At the same time, however, the Italian AI research ecosystem presents four main weaknesses this strategy will address, and namely:

• Parceled growth of research labs. Despite the international excellence of the Italian AI research community, the persistent issue of scale, critical mass and interdisciplinary integration of research labs, together with the scarce public/private national funding of fundamental research, represent a serious potential risk for the country’s competitiveness.
• Poor talent attraction. While new talents’ development is adequate in Italy, talent attraction is poor, with few foreign talents coming to work in Italy. No special strategies have been effective, until now, to support the attraction of foreign talents.
• Significant gender gap. A survey of 2020 of CINILab AIIS National Assembly among its members shows that only 19.6% of AI researchers are women, compared to close to 50% in STEMS overall. [4].
• Limited patent capacity. Italy yields far less patents per paper than EU peers.

2. Knowledge transfer centres

Alongside universities and research nodes, Italy has a wide network of knowledge transfer centers, promoted by industrial associations and institutions at the national and local level. Among these, it is worth mentioning 20+ Digital Innovation Hubs (DIHs) promoted by Confindustria [5] , 70+ Punti Impresa Digitale promoted by Italian Chambers of Commerce [6] and 8 competence centres [7] promoted by the Ministry of Economic Development. These centres, which belong to the Network Industria 4.0, offer services ranging from sensibilization and awareness campaigns, to technology assessment, strategic roadmapping and employee training. In addition to these national resources, Italy is active in the EU-DIHs network [8], which will function as one-stop shops that help companies dynamically respond to the digital challenges and become more competitive.

3. Technology and solution providers

The Italian ecosystem is characterised by a variety of technology providers which specialise in a variety of AI solutions. As of February 2021, the Observatory on AI [9] registered 260 Italian companies offering AI products and services (See Figure 1):

• The majority of them (55%) offer solutions designed to be used in specific areas, i.e. Healthcare, Marketing & Sales, Finance and Cybersecurity.
• One quarter (25%) are more general in nature, providing for example advanced analysis of structured and unstructured data for purposes that may concern Forecasting, Classification, Natural Language Processing, Computer Vision, Human-AI interaction in various application areas. About a third of these companies are start-ups [10] .
• The remaining companies are divided between System integrators (about 10%) and Consulting companies (about 5%), most of which are either medium or large companies.
• Few of them are cloud providers offering AI solutions and companies offering sensing and edge computing for efficient computing or storage dedicated to AI. Cloud providers are limited in number but they tend to be among the largest players. The largest among them offer services in critical infrastructures management or cybersecurity.

Figure 1: Italian companies offering AI products, by sector

In addition, several large corporations are creating in-house AI research labs to create solutions and products that may cater to their business, their clients or both. Companies are most active in the energy, security, aerospace, telecommunications, insurance, banking, cloud and household and personal care (HPC) sectors.

Lastly, the Italian entrepreneurial ecosystem registers a growing number of start-ups with AI expertise; CINI Lab AIIS in 2020 surveyed 110 university spin-offs or start-ups connected with research centres, focusing on machine or deep learning, big data analytics and human-AI interaction. This figure is particularly relevant as AI is one of the sectors where startups concentrate the most, showing a national proclivity for the sector.

Despite the considerable number of players / start-ups, the AI market in Italy is still very limited in size and this is negatively affecting the growth of our companies and their investment capability. Considering the fast development rate of AI, this represents a weakness that should be addressed.

4. Private and public user organisations

Lastly, the Italian ecosystem is composed of private and public users organisations. As for private organisations, in 2020 according to a research conducted by the School of Management of Politecnico di Milano, 53% of medium-large Italian companies declare that they have started at least one AI project. The sectors that show the greatest diffusion of fully operational projects are manufacturing (22% of total projects started), banking-finance (16%) and insurance (10%).

Table 2: AI projects within Italian medium-large companies, by sector

In 2020, the Artificial Intelligence private market in Italy reached a value of €300 million [11], an increase of 15% over 2019 but that still places Italy’s contribution to Europe’s AI market below its GDP contribution (ca. 3% vs. ca. 12%) [12]. Within the Italian market, €230 million (77%) was commissioned by Italian companies, while €70 million (23%) was project exports. The Politecnico study highlights the following classes of AI based solutions:

• Intelligence Data Processing (33% of the expenditure),
• Natural Language Processing (18%),
• Recommendation Systems (18%),
• Intelligent RPA, Chatbot / Virtual Assistant and Computer Vision (31%).

With regards to the public sector, understanding what is happening in the field of digital transformation is crucial in order to imagine how AI can create synergies with the process of digitising the country. According to the study E-government Benchmark 2020 by the European Commission, [13], Italy is characterised by a low level of penetration (25% vs. 60% of EU27+UK) and a medium level of digitisation (71% vs. 72% EU 27+UK). Therefore, there is still a gap between the supply of digital services and their actual use. Although the digitisation level is in line with the performance of other European countries with similar environments, online services are not fully widespread all over the country.

Italy’s international position¶

The Italian AI ecosystem is in line with international peers [14] when it comes to research quality and output. Yet, this ecosystem fares less well when it comes to business spending on R&D, patenting and AI applications. Table 3 summarises key statistics.

Table 3: Comparison and international position

National R&D resources					Learn more
R&D expenses (as a % of GPD, 2019)	3,17%	2,19%	1,76%	1,45%
Share of GDP in Research [15] (as a % of GPD, 2019)	0,46%	0,28%	0,13%	0,20%
R&S (€Mln, 2019)	109.544 €	53.158 €	44.364 €	25.910 €	Learn more
R&D personell per millions inhabitants (2018)	8.500	6.950	7.000	5.150	Learn more
AI research and patents statistics
AI pubblications (2019)	2.660	2.755	2.974	739	Learn more
AI pubblications (2019)	5.310	3.352	6.645	3.374	Learn more
Average AI researchers productivity [16]	2,00	1,22	2,23	4,57
Patenting strategy intensity index (%global patents / %global publications)	0,79	0,34	0,29	0,07
Patents Equivalent patent applications 2019	178.184	67.294	54.762	32.001	Learn more
Number of agents [17]	147	76	163	42	Learn more
Business investments and outcomes on AI
Business spending on R&D (€Mln, 2018)	74.162 €	33.809 €	28.926 €	14.691 €	Learn more
R&D business spending on R&D (% of the PIL, 2018) [18]	2,23%	1,45%	1,22%	0,84%	Learn more
Number of Global Digital Champions [19]	8	7	33	0
Government AI-dedicated declared investments (€ Mld)	5,0 €	3,0 €	n.a.	Output plan	Learn more

Research position

Italian AI researchers have a widely recognized international standing regarding scientific output, both in terms of quantity and quality [20].

Nevertheless, there are several limitations affecting the Italian research landscape - both in the public and private sphere - which need to be explicitly considered and addressed. In particular:

• The public research system is less funded than peers and wages are low. On average, EU countries (EU28) [21] invest 2.38% of GDP in research, while Italy invests 1.45% of GDP only. Low funding is compounded by low wages, as public research institutions must comply with the same rules as the rest of the Public Administration, without considering the specific needs of these entities. As a consequence, Italian researchers receive lower wages (€15,343 for PhDs) than in Germany (€48,027), France (€20,220), UK (€17,272 tax-free) and Switzerland (€67,998). [22]
• In the private sector, firms are underinvesting in R&D. Although Italian companies spend €14.7 billion in R&D per annum (2018) this figure is less than average EU peers. Investments are also constrained by the small size of the average Italian firm. Lastly, Italy is missing global Digital Champions [23] in the hardware, software and integration areas, and is thus missing a key factor in stimulating innovation.

Technology & solution providers

The Italian Market for AI solutions is still small, as Italian companies are lagging in the digital transformation process. The adoption rate of AI solutions by Italian corporates is lower (35%) than the EU average (43%) and in 2017 only 15% had projects beyond the pilot phase (vs. 32% for European peers) [24]. Italian corporations indicate the high costs of adoption and the absence of public funding as the main obstacle. Moreover, Italian companies struggle to find qualified AI professional [25]. Since AI is a significant enabler of productivity growth, this late adoption will negatively impact GDP growth, widening the gap between Italy and peer countries. According to McKinsey [26], while AI-driven GDP will grow at 19% for the EU , it will grow only by 13% in Italy.

Summary: Challenges ahead

The evidence provided so far calls for a radical upgrade and update of Italy’s national AI strategy. As described, the country needs to enhance the positive elements of its current ecosystem while concentrating on reforming and investing in specific areas of weakness. In short, Italy needs to:

• Strengthen its AI research base and funding,
• Foster measures to retain and attract talent,
• Improve its technology transfer process,
• Increase AI adoption among firms and public administration as well as foster the creation of innovative enter- prises.

The remainder of this document will provide the details of Italy’s plan going forward.

The first two cycles of PhD-AI.it have been co-funded with about €8 million by the Ministry of University and Research and with the same amount by the universities and research organizations participating in the programme.

The anchoring: an EU centred AI¶

Over the last decade, industry and society have been radically changed by the ongoing digital revolution. In this context, the EU has recognized AI as an area of strategic importance with the potential of becoming a key driver of economic development.

In April 2018, the European Commission put forward a European strategy on AI in its Communication ‘Artificial Intelligence for Europe’ COM(2018)237 a s part of its Digital Single Market Strategy. This strategy delineates the following aims:

• Boosting the EU’s AI technological and industrial leadership and AI uptake across the economy, both in the private and public sectors.

• Preparing for socio-economic changes brought about by AI.

• Establishing an appropriate legal framework based on ethical values and principles.

Following this, the European Commission and the Member States put forward a ‘Coordinated Plan on Artificial Intelligence’ - COM(2018)795 in December 2018 with the stated goal of maximizing AI investments’ impact both at European and national levels and strengthening synergies and cooperation among Member States. To this end, Member States were strongly encouraged to develop their own national AI strategies as a first step to achieve these aims. This strategic programme responds to this encouragement.

More recently, the EU Commission published the ‘2021 review of the Coordinated Plan on Artificial Intelligence’ with a view to strengthen Member States’ competitive landscapes. The review reports actions taken by the European Commission so far within the framework of the MFF 2021-2027 programming and sets concrete proposals and recommendations for further joint actions between the EU and Member States. The plan was accompanied by significant investments: €1.5 billion (2018-2020) and €1 billion/year (after 2020) to be allocated in the following funding framework programmes [1].

• Horizon Europe

  • Aspects contributing to AI (data and cloud infrastructures for AI). (data and cloud infrastructures for AI)).

  • Development of AI technologies (especially in cluster 4, where the partnership AI-Big Data and Robotics has been launched, and bottom-up approaches like EIC, MSCA, ERC etc.). :name: Development-of-AI-technologies-especially-in-cluster-4-where-the-partnership-AI-Big-Data-and-Robotics-has-been-launched-and-bottom-up-approaches-like-EIC-MSCA-ERC-etc.).

  • Use of AI in other clusters contributing to AI (1. Health, 3. Civil security for society, 5. Climate, energy and mobility, 6. Food and natural resources) and in all Missions. :name: Use-of-AI-in-other-clusters-contributing-to-AI-1-Health-3-Civil-security-for-society-5-Climate-energy-and-mobility-6-Food-and-natural-resources-and-in-all-Missions.

• Digital Europe

  • Strategic Objective 2.

  • AI-related activities in SO4 (skills) and SO5 (European Digital Innovation Hub).

To support the implementation of the plan outlined above, the European Commission launched the project “AI Watch” in 2019, with the goal of monitoring industrial, technological and research capacity, as well as national policy initiatives. Furthermore, the project keeps track of updates and technological developments of AI worldwide, assesses their impact in the economy, society and public services. Its first national strategies’ review was published in February 2020.

The last step taken by the European Commission, in April 2021, is the launch of the proposal for a European Regulation on AI [2] which aims to introduce a common regulatory and legal framework based on a well-defined risk-based assessment of each field of application.

The initiative also proposes the introduction of a European Artificial Intelligence Board, which will encourage national cooperation and ensure that the regulation is respected.

AI is therefore a key priority for the European Union. Italy aims to be aligned with this European priority and integrate its Strategic Programme within this European approach.

Guiding Principles¶

This National Strategic Programme for Artificial Intelligence defines a coherent and holistic framework of initiatives to support the development of a national AI ecosystem. Its design is inspired by five guiding principles.

1. Italy’s AI is a European AI. In line with the EU Coordinated Plan on Artificial Intelligence [3], the Italian Strategic Programme stems from the awareness that only through common and coordinated actions Europe will be able to compete globally and work towards strategic autonomy. Therefore, this programme reflects the four sets of proposals put forward by the 2021 EU Coordinated Plan on AI. First, it sets enabling conditions for AI’s development and uptake by focusing on cooperation, data and computing infrastructure. Second, it leverages on existing Italian HPC and data-management infrastructure. Third, it aims at nurturing talents and adheres to the joint effort for improving and adopting the harmonised set of rules for AI proposed by the AI ACT. Fourth, it identifies priority sectors where to build strategic leadership.
2. Italy will be a global research and innovation hub of AI. [4] To guarantee future economic growth and strategic autonomy, it is essential for Italy to boost its AI research and development ecosystem and leapfrog at the forefront of AI developments. Accordingly, this strategic programme will invest in frontier research and applications to develop AI methodologies and techniques of tomorrow.
3. Italy’s AI will be human-centred, trustworthy and sustainable. [5] Technologies must not promote economic growth per se, but inclusive and sustainable growth, in line with the principles contained in Article 3 of the Italian Constitution and the United Nations Sustainable Development Goals. This means that AI development must be centred around economic and social inclusion, human rights as well as environmental sustainability. AI must be designed and implemented in a responsible and transparent manner, based on trust and robustness so that it can be safely adopted in every sector and be capable of responding to societal challenges. To this aim, Italy adheres to the “Ethics Guidelines for trustworthy AI- Guidance and implementation program” defined by the High Level Expert Group on AI.
4. Italian companies will become leaders of AI-based research, development and innovation. The digital transformation of our entrepreneurial ecosystem is a must, if Italy wants to keep up with the most developed and innovative nations. To that end, this programme fosters the development, implementation and adoption of AI solutions. Public-private partnerships will be instrumental in finding appropriate synergies between research bodies and enterprises with the aim of increasing Italy’s technology transfer capabilities and thus competitiveness.
5. Italy’s public administrations will govern with AI and will govern AI. [6] The use and impact of AI in the public sector revolves around the dual dimensions of governance ‘with and of’ AI. On the one hand, Italy’s Government will improve its internal processes and policies thanks to a responsible use of data and AI technology. On the other hand, the Government is committed to governing AI and mitigating its potential risks, especially to safeguard human rights and ensure an ethical deployment of AI.

These Guiding Principles are not separate pillars. They offer an ambitious vision of the future of AI in Italy and an integrated approach that aims at stimulating all those factors that can contribute to the development of a favourable ecosystem. Industrial changes will lead to societal ones and both of these, in turn, will stimulate fundamental and challenge-driven research. The Strategic Programme is set to support all of these aspects.

Objectives and Priority Sectors¶

Given the challenges laid out in the prior chapter, as well as Italy’s competitive position in the field of AI, this strategic programme lays out six objectives with a view to boosting Italy’s strengths while mitigating its weaknesses. These objectives are:

Objective 1: Advance frontier research in AI, both fundamental and applied, with a view to generating impact on important priority sectors of Industry, Public sector, Society, and Environment. A multidisciplinary approach must be incentivized, where research coexists with industrial and social innovation generating true innovation ecosystems.

Objective 2: Reduce AI research fragmentation by helping the AI ecosystems achieve critical mass and by fostering network collaborations, making scientific excellence coexist with social cohesion and territorial inclusiveness.

Objective 3: Develop and adopt human-centred and trustworthy AI in the public and private sector, promoting the societal acceptability of AI solutions, their compliance to regulation of AI, while supporting the development and design of responsible AI technology and systems.

Objective 4: Increase AI-based innovation and the development of AI technology by fostering industrial investments and partnerships that drive excellent science out to the market, hence facilitating the uptake of AI in SMBs.

Obiettivo 5: Develop AI-driven policies and services in the public sector by boosting public sector innovation, the adoption of AI solutions and the cooperation between research centres, industries, and public bodies.

Obiettivo 6: Create, retain and attract AI talent in Italy by promoting all levels of education in AI, creating a new generation of holistic AI researchers and innovators, making Italy an attractive AI destination for qualified human capital from abroad, with a particular attention to diversity and gender balance.

Together with this strategy’s six objectives, Italy is committed to investing in eleven priority sectors. These sectors include those where Italy already has a competitive advantage, such as manufacturing, culture, agri-food and health, but also industries that are strategic to the country’s technological development, such as national security, IT and the environment.

Industry and Manufacturing. AI will enable the Italian manufacturing sector, the second largest in Europe and seventh largest in the world, to introduce innovative processes, products and business models to the market while maintaining (or increasing) its global competitive edge.

Education system. As artificial intelligence is transforming every aspect of our lives we need to educate all people to this technology through a new education and training plan to understand, reinforce, integrate and disseminate AI technology. AI should be an important topic at all education levels. At the same time, it can constitute a powerful instrument for a fruitful transformation of the national education system to develop personalized learning plans while ensuring fairness and trustworthiness.

Agri-food. Through AI, Italy’s thriving agri-food sector has the potential to increase further its competitive position by developing precision agriculture, thus avoiding overproduction and waste, increasing food safety and reducing emissions from land and agriculture.

Culture and tourism. Advanced technologies will further increase Italy’s touristic attractiveness by creating new synergies between cultural and creative industries, producers, managers and users of Italy’s vast cultural heritage. These technologies enable, for instance, continuous monitoring and preventive restoration of cultural heritage, monitoring and alert system for landscape heritage, customisation of services to better meet demand, virtual tours of tourist destinations to allow for better informed choices of travel destination, simultaneous translators for the description of places and monuments visited, geolocalised services for tourists.

Health and wellbeing. In the field of healthcare, AI applications boost product and process innovation by exchanging and aggregating information that is currently scattered in a multitude of public and largely underused databases. AI applications will help meet the new needs arising from an ageing Italian population. Moreover, they will have a significant impact on the population at risk of severe diseases such as degenerative, oncological, and viral diseases, and increase social inclusion of disadvantaged groups. A few application examples are medical devices and services in screening and diagnostic areas such as omics and medical imaging, new drugs and vaccines, tracking and treating people, supporting patient care (diagnosis and prognosis), and predictive models of healthcare needs.

Environment, infrastructures and networks. AI solutions will have a significant impact on preserving resources, reducing emissions, better managing traffic flows and related risks, strengthening the circular economy and better preventing natural disasters. More generally, AI will be a fundamental ally in accelerating the ecological transition, a pillar of Italy’s recovery and resilience plan and European Union’s recovery efforts. In addition, AI may have an enabling role also in the highly strategic development of 5G networks as it can help improve network performance as well as reduce capital expenditures associated with its infrastructure deployment/management [7]. A few application examples are monitoring and intelligent management of networks and consumption, monitoring and predictive management of the waste cycle, situational and predictive analysis of hydrogeological instability.

Banking, Finance, and Insurance. Modern AI technologies will allow banks and insurers to improve in at least two ways. First, they will increase the quality of services offered to customers and reduce their costs through a higher level of personalisation and security of transactions. Second, AI applications will strengthen fraud prevention systems and simplify the fulfilment of intermediary obligations through the adoption of mechanisms for detecting suspicious behaviour and analysing data and documents.

Public Administration. In the near future, AI will optimise bureaucratic processes, offering better services to citizens and businesses and reducing costs [8] for better services and performance. Furthermore, with its databases and innovative tools for purchasing, investment and regulation, the PA is called upon to play an active role in the AI revolution in the private sector for the benefit of the community (e.g. open data, geolocation tools, purchase of AI products and services, funds for AI, experimentation of AI solutions). The PA could benefit from AI solutions concerning flow management, virtual assistants and chatbots, predictive analysis of business risks and support for the examination of incentive applications, support in the fight against tax evasion and other forms of illegality, evaluation of past policies and impact analysis of experiments.

Smart cities, areas and communities. The COVID-19 pandemic has shown that the digital ecosystem is essential to support all citizens, whether they live in cities or rural areas. AI will enable Italian residents, wherever they live, to gain access to communities and services, while reducing costs. Finally, AI technologies will enable Italy to reduce traffic and limit congestion thus also contributing to reining in the effects of one of the most polluting activities in the country. A few examples are smart parking, traffic management and signage control, self-driving vehicle management systems, lighting management and optimisation of public transport, as well as monitoring of bridges and buildings, home automation for buildings.

National Security. The importance of AI for the National Security of a country has been growing steadily in the last five years. Hence, Italy is fully committed to investing in AI applications that ensure the security of its citizens. This includes individual and national cybersecurity, where AI has been contributing to the development of new-generation detection and resolution software.

Information Technologies. The success of applications of AI in the sectors described above strongly depends on a high level of innovations in IT crucial fields impacting AI, such as Sensing, Reasoning and Search, Natural Language Processing, Computer Vision, Human-AI interaction, and Edge Computing. The broad field of IT [9] has a crucial role in ensuring a high level of innovation for implementing competitive AI in all different applications. For this reason, a special effort will be devoted to supporting the birth and growth of Italian IT companies.

Talent and skills¶

AI has a transformative impact on all aspects of our society, and the COVID-19 pandemic has accelerated this trend. The 2020 World Economic Forum estimates that by 2025 85 million jobs may be displaced by a shift in the division of labour between humans and machines, while 97 million new roles may emerge. This transition will have significant distributional effects both across and within countries.

To mitigate the effects of such a transition (e.g., unemployment in certain segments of the population) and at the same time ensure that Italy remains on the edge of the technological frontier, the country needs to invest holistically in AI training and skills development for citizens. On the one hand, therefore, the country needs to invest in being at the forefront of AI research. This means expanding and improving the PhD programmes and attracting/retaining top researchers. On the other, Italy must ensure that the whole economy takes advantage of the productivity opportunities inherent in the diffusion of AI. Thus, it needs to strengthen the STEM component of the broader educational system, in order to support the development of a workforce that can interact with AI and exploit its benefits.

Research¶

As shown in chapter 1, the Italian research ecosystem shows signs of strength, yet its fragmentation, lack of resources and low patenting rhythm impede its effectiveness. To address these challenges, this section illustrates policies that aim to bridge the gap between foundational and applied research by fostering collaborations between academia, industry, public bodies and society. The future of AI necessarily implies a strong synergy among public and private research centres, industrial research and innovation centres, start-ups and SMEs, and target users’ domain expertise.

Research programme initiatives are divided into two classes:

B. Building the Italian AI research ecosystem: fundamental and applied research.

C. Horizontal aspects.

Applications¶

As evidenced in the introductory chapters, the Italian AI ecosystem suffers from low patenting and a slow technology transfer process. In addition, Italian firms, large and small, have so far been slow to adopt AI solutions resulting in an AI market of limited size.

To address these challenges, this strategy proposes a set of policies aimed at broadening the breadth of AI application in industries and society, as well as measures to foster the birth and growth of innovative AI enterprises. In addition, these policies are meant to insist on priority areas and accompany the growth of sectors that have so far shown potential in AI development and adoption.

All initiatives share common issues and targets:

• To pay particular attention to smaller companies, those operating in the most peripheral and disadvantaged geographical or socioeconomic contexts, focusing on the priority sectors (Section 2.3) and on national strategic sectors (Critical Infrastructures, sectors defined in “Decreto Golden Power”).
• To increase the number of female AI entrepreneurs and experts, as well as attracting foreign AI-based start-ups and practitioners with economic incentives to be applied in all of the initiatives outlined below.
• To align all AI policies related to data processing, aggregation, sharing and exchange, as well as data security with the National Strategy for Cloud and with the initiatives underway at EU level, starting with the European Data Strategy and the recent proposal for a Data Governance Act and AI Act.

To that end, this strategy identifies two areas of intervention that we deem to be of highest impact as well as most strategic.

D. AI for more modern enterprises.

E. AI for a more modern public administration.

D.AI for more modern enterprises.

The impact of AI on businesses will be of enormous relevance and should concern all enterprises. In fact, AI implies a real revolution in their modus operandi, from internal processes and customer relations to the development of new AI-based products and services. In turn, AI implies that Italian corporates would need to transform their workforce as well as processes, hiring new talent, upskilling the existing workforce and making sure such transition is carried out with the most effective and responsible use of AI solutions.

Overall, the proposed initiatives aim at:

• Supporting the hiring process of highly skilled AI personnel in private companies, so as to reinforce their 4.0 Transitions process (machinery, HW, SW and people),
• Increasing the adoption of AI solutions in private companies, so as to increase their competitiveness,
• Helping start-ups and spin-offs to scale up, avoid the “valley of death” and support their national and international growth,
• Establishing a regulatory context that may help the experimentation and the certification of reliable AI products and services that have passed such experimentation.

Therefore, this strategy supports the following initiatives: