Artificial Intelligence: the expert task force releases its final proposals for the Italian strategy
Putting the citizen and society at the centre of AI
10 September 2020
In July 2020, a task force of thirty experts, entrusted by the Italian Government to draw up a "proposal for an Italian strategy for artificial intelligence" (the "Task Force"), released their final report (the "Report"), after more than two years of work.
In the 115-page long document, the Task Force sets out 82 proposals relating to the development of AI in Italy, tailored to the peculiarities of the Italian system and aimed at combining international competitiveness and sustainable development.
As confirmed by the Task Force, their work has been driven by a clear anthropocentric matrix and oriented towards sustainable development.
The Report is structured in three parts: the first part is dedicated to the analysis of the global, European and national market of artificial intelligence; the second part describes the fundamental elements of the Italian strategy; the third one sets out some recommendations for the implementation, monitoring and communication of the national strategy on artificial intelligence.
AI's trends from a global, European and Italian perspective
The development of AI has become a central theme for both industrialised and emerging economies in recent years. The two key players on a global scale are the US and China. In particular, in terms of investment, the efforts currently being made by the two superpowers in the AI domain is far greater than that of any other country.
China has clearly stated its aspiration to become the world leader in AI by 2030 and has put in place a number of ambitious state lead initiatives to reach this goal. The United States have adopted a less managerial approach to AI development, partly due to the preponderance of the private sector in R&D spending in this field. In May 2018, however, the White House announced its ambition to maintain leadership in AI, combining market objectives with the need to protect employment and promote publicly funded research and development.
Alongside the two superpowers, industrialized countries such as Japan, South Korea, Canada and emerging economies such as India have adopted national AI plans.
The EU institutions have not been unreceptive to the proliferation of national plans regarding AI and have decided to strengthen coordination between Member States within the "Digitalising European Industry" program. This effort resulted in the "Coordinated Plan on Artificial Intelligence", adopted on 7 December 2018.
The Coordinated Plan recognises the need for stronger coordination amongst Member States to become the world-leading region for developing and deploying cutting-edge, ethical and secure AI and it proposes joint actions for closer and more efficient cooperation in four key areas: (i) increasing investment, (ii) making more data available, (iii) fostering talent and (iv) ensuring trust. The document also represents that, compared to other parts of the world (such as the US and China), investments in the sector in the EU are still modest and fragmented, therefore calling for greater coordination of investment amongst Member States.
The Coordinated Plan is accompanied by two sets of initiatives, one at the EU level and the other at the Member States level. At the EU level, the European Commission has set up a High Level Expert Group to support the EU strategy with two key inputs: (i) the definition of guidelines on AI ethics, and (ii) the formulation of recommendations on public policies and investments to be made in Europe to promote the EU's competitiveness in the field of AI. In parallel, Member States have committed themselves to defining national AI strategies.
The Italian Report points out that a first look at those national strategies reveals a marked emphasis on the need to strengthen the competitiveness of the country in what is defined by many as the most important technological evolution of the coming decades. The plans available so far also look at measures aimed at facilitating the availability of data for research and innovation, the promotion of skills, and the creation of administrative structures to enable the evolution of technology and its use in public administrations and among citizens. However, so far, no country has placed sustainable development at the centre of its strategic plan.
As far as Italy is concerned, the Report recognises that the spread of AI services in the country is still limited: it is estimated that only 12% of Italian companies have activated a project in this field, the most widespread of which concerning virtual assistants. Nevertheless, Italy, like many other countries, is laying the foundations of the new AI ecosystem and the growth of the AI market is relevant (50% per year).
According to the Task Force, the following are the key areas of expertise in the AI sector that Italy should leverage on and further develop:
Embedded AI: Italy has a strong specialisation in the combination of AI with physical systems that include sensors, intelligent objects, robots, automation systems, etc. Embedded AI is certainly a distinctive factor of Italian know-how in this field. In addition, Italy is certainly a top player in the development of intelligent hardware and software components and IoT objects, often combining technology with design and Made in Italy;
Automation industry: Italy is at the top European level in the industrial manufacturing and automation sector. Italian robotics has grown on average 12% in the last five years with a peak of 19% in 2018. Important growth is also expected in service robotics, in a world market that already exceeds USD 11 billion;
Customer care: Italy has extensive expertise also in the application of AI in the customer care industry and, more in general, in the customer relationship management (marketing) on digital channels that use natural language processing. The use of machine vision for user authentication is pervasive;
Cultural heritage: in the cultural heritage sector, Italy, for its heritage and skills, is at the top of the world. This is also true for the technological offerings for the use, preservation and experiential interfaces, as well as for the protection and preservation of cultural heritage. In particular, many start-ups and SMEs in the cultural and creative industry develop AI solutions in the context of augmented reality and human-machine interactions.
The Task Force invites the Italian Government to claim an important role in future research centres at European level. In particular, if the European Union were to set up a research body spread over several poles, Italy should apply for a leading role in robotics and embedded AI.
The Italian Strategy: the pursue of a anthropocentric, trustworthy and sustainable AI
The Report calls for the legislator to pursue not only industrial competitiveness in the AI industry, but to focus its efforts on the use of AI to improve the welfare of citizens, and to implement AI in order to minimise the risks for the individual and for social cohesion.
Specifically, the Task Force urges for the Italian AI strategy to take a step forward compared to what has been outlined so far by other countries and the European Union, by putting the human being and the society as a whole at the centre of the AI strategy. This implies a series of careful and forward-looking policies to ensure that:
- AI is fully in line with existing legislation and fundamental rights;
- the precautionary principle is fully adopted in the introduction and design of AI systems, combined with a careful approach to the needs of innovation;
- forms of AI that increase human intelligence, productivity and creativity, rather than replacing them, are encouraged, also with significant investments in education;
- AI fully respects the personal integrity of the individual, including a set of transparency obligations and consumer/user rights; and
- in attributing liability for damage caused by AI, the damaged party can be fully and timely compensated for his/her loss.
The Task Force also identifies the three key pillars for the Italian AI strategy:
The first pillar concerns the individual and his/her relationship with "the machine." In this regard, the Report stresses the need for the Italian AI strategy to put particular emphasis on the ability for the individual to self-determine him-/herself, to preserve his/her privacy and to have an adequate education, support and protection in the use of AI systems as an individual, as citizen, as user, as consumer and as worker.
For this to be achieved, it is necessary that citizens are made aware of and understand the potential and limits of AI. Therefore, on the one hand AI systems must be, where possible, fully transparent and explainable to the user without advanced scientific and technological knowledge and, on the other hand, citizens must have the necessary skills and knowledge. The legislator should therefore encourage educational programs, both in lower and higher education and lifelong learning programs that will provide the necessary soft and hard skills required to understand and interact with AI.
More generally, the Government should plan for popularising initiatives concerning AI, designed by experts in the field and embedded in an organic, organised and strategic framework. These initiatives should have different targets, such as businesses, the public administration, the service sector and society as a whole, in order to achieve a positive and long-term impact.
In addition, an anthropocentric approach to artificial intelligence cannot ignore the protection of the individual as consumer and user. Protecting the consumer-user undoubtedly means completing the framework of compensation for damages suffered due to the use of artificial intelligence systems. But a real protection of the user must also take into account other aspects, more specifically related to the context of the new AI-based services. In this regard, the European Commission has already launched a review of the current legislation on civil liability.
Finally, attention should also be paid on the impacts of AI on businesses and the job market. In particular, it is undeniable that with the expansion of AI technologies, some repetitive tasks, which require low rate of professionalisation and competence, will be replaced by algorithms; however, at the same time other jobs may be created that will require new skills. The Report stresses that it will be essential to invest in training and reforming the entire companies' training system to guarantee to employees the access to long life learning programs and the possibility for a re-qualification.
The second pillar concerns the need for a reliable, productive and sustainable digital ecosystem.
According to the Report, to lay the foundations of a trustworthy AI it is necessary to adopt a broader vision than just aligning AI systems to ethical principles. Indeed, it will be necessary to include requirements that will ensure the safety and robustness of the technologies employed. In the AI system, the concept of "safety" embraces a much broader perspective than the purely individual one, extending also to micro and macroeconomic aspects, to employment, to social security, and to environmental aspects (in terms of sustainable development).
Therefore, the Report calls for the drafting of a document that refers to specific areas (not only ethical or legal, but also purely technical/engineering) affected by AI, with an indication of "minimum requirements" (in terms of compliance with ethical principles, legal standards, self-assessment of risks and technical requirements) that each actor involved must ensure to operate through an AI system. This could be made in the form of a "checklist" which lists the ethical principles to be complied with, as well as the regulatory requirements needed, the minimum technical measures to be taken and the self-assessment of the main risks and mitigation measures.
In addition, the Task Force also points out that the rise of AI as a tool for optimisation, data processing and forecasting is an unprecedented opportunity for a reform of the Public Administration. This revolution can be carried out by employing re-engineering processes, re-use of data and interoperability between administrations, in order to allow the individual user to easily access public services, to maintain control of his/her personal data and to benefit from the increased exchange of data between administrations. Furthermore, the Public Administration should take this opportunity to implement data-drive techniques in policy-making.
The third pillar includes actions related to environmental protection and sustainable development. As the Report points out, Italy, like the rest of the world, faces enormous challenges over the next few years, which will require radical changes in the way we produce and consume: AI, with its complementary and enabling technologies, can play an essential role in this transition.
The Task Force invites the Italy to be an ambassador and a creator of AI solutions for sustainable development, in order to address the most pressing social and environmental issues and, at the same time, act as a promoter of technological solutions in other European countries and in the world, particularly in developing countries.
The two main sectors in which AI solutions could be implemented are the following:
- Sustainable energies: AI technologies can be effectively used to improve energy production, management and distribution. In fact, new energy ecosystems will be based on more just-in-time and smarter paradigms, and with a high dependence on data and technology. A good example of this is provided by the offshore wind sector, where, in a few years, the market has been transformed into a totally new structure, bringing more competition and more supply also due to the use of AI technologies.
- Accessibility and social inclusion: AI technologies can also be implemented to improve the ICT accessibility aspects for individuals with disabilities, or increasing the inclusion of disadvantaged categories of citizens. In particular, the Task Force urges for the research to focus on developing services and products to ensure better access, including an advanced multimedia visual and sound experience, to people with disabilities, by providing to advanced forms of service customisation. In addition, the development of semantic technologies for the extraction of digital content, including web content can make navigation and access easier and more effective for visually impaired people. Also, image and gesture recognition technologies can improve the understanding and even translation of sign language.
The Report's key findings and its recommendations
Overall, the Task Force envisages a concrete possibility for a significant cultural, economic and social growth attributable to an anthropocentric, trustworthy and sustainable AI – what the Report defines "a new reinAIssance."
According to the Report, this approach is far from being purely ideal: posing the problem of the development of AI from the point of view of the future of work, interpersonal relations and social and environmental protection is a Copernican revolution compared to the traditional approach oriented towards industrial competitiveness, which can play an essential role for the sustainability and development of our economy and society.
Nevertheless, the Task Force warns that this will be possible only if we will be able to think, develop, and regulate the future relationships between man and intelligent machines from both a regulatory, ethical and political point of view, in order to preserve the pivotal role of the human being, in his cultural and material wealth. Italy should be an example of this approach to AI at national level and an ambassador in the European and global context.