For those who do not have time: Summary of the White Paper

This White Paper aims at analysing the impact of Artificial Intelligence (AI) in our society and, more specifically, in the Public Administration, in order to promote digital transformation. It is the result of the work of the Agency for Digital Italy and with a task force of experts.

The objective is to facilitate the adoption of these technologies in the Italian Public Administration, to improve services to citizens and businesses, thus giving a decisive impulse to innovation, the proper functioning of the economy and, more generally, to progress in daily life.

Today, Artificial Intelligence can drive vehicles, take care of elderly or sick people, perform dangerous or weary jobs and help make informed decisions based on the rational management of large amounts of data. Also, AI can enable us to communicate in languages we do not know, help us study and increase the cultural or entertainment experiences at our disposal. In the Public Administration, it can be used profitably in the healthcare, education and judiciary systems, public employment, security and, more broadly, in the management of relations with citizens, which can be simplified and at the same time be more effective, quick, and efficient.

As prerequisite to achieve these objectives is necessary to define “Artificial Intelligence”, from both a theoretical and a technical point of view. This serves to understand opportunities and limits of these technologies, as well identify their most effective areas of application. The White Paper focuses on tools that can have positive effects on the work of the Public Administration, such as the use of chatbots to answer citizens’ queries and cut through layers of bureaucracy, or the use of robots to take care of the sick, algorithms that can read results of medical exams and these that support the preparation of educational path of students in order to improve their performance. Particular emphasis is placed on the means for monitoring and managing careers, those for the surveillance of public places or for the recognition of network threats, as well on tools for the rational management of problems generated by natural disasters.

The successful implementation of these technologies in public services requires the public sector to evaluate the critical challenges that must be faced to integrate Artificial Intelligence in a profitable way.

The first challenge is the Ethical challenge, which remains central to any consideration in this field. There is a need to strongly affirm the anthropocentric principle stating that Artificial Intelligence is always at the service of the citizen and not vice versa.

Moreover, it is necessary to formulate general principles of equity with the aim of using these new technologies to address some universal needs such as to respect freedom, and to guarantee individual and collective rights. More specifically, in the field of ethics, the functioning of Artificial Intelligence raises some issues, i.e. those related to the quality and neutrality of data, the responsibility of those who use algorithms, transparency and accountability, as well as protection of privacy.

The aim is to demonstrate how the incorrect use of the technologies at our disposal can contribute to the development of a more unjust society that fuels inequalities, while awareness of the risks involved in relying on “smart” machines can help us minimize inequalities and plan a better world.

The second challenge that is addressed is the technological one: Artificial Intelligence is not yet able to reproduce the complex functioning of the human mind, but only some of its circumscribed capacities. Therefore, one of the goals is to make existing and future technologies more similar to our way of relating to the world.

In the short term, the priority is to work on personalization and adaptivity to ensure that data and algorithms at our disposal can be increasingly more effective in allowing us to operate individually in some specific areas of our daily life. For the purpose of this White Paper, the focus is on rendering more effective the work of the Public Administration (in addition to the sectors mentioned above: tax, mobility and transport).

The document then analyzes the fundamental issue of skills that must be developed in the age of Artificial Intelligence. The issue is addressed from both the citizens’ and the civil servants’ points of view. Citizens in general, and primarily those who would like to operate in the IT sector, need to understand how algorithms and databases that are used by the Public Administration, civil servants need to constantly ameliorate their skills.

However, even for those who want to engage in other sectors, it will be necessary to know how to act in contexts that will increasingly assign to machines tasks that were previously performed by people (typically, the simplest and most repetitive ones). This challenge underlines the importance of understanding how to relate with machines and coexist with them in a world that is always more populated by technology. The State bears the responsibility to respond to these needs by both shaping an education system able to keep up with these changes and encouraging lifelong learning. However, in order to achieve these goals, civil servants and people employed in the public sector must be properly trained. They will face the need to understand what are the Artificial Intelligence tools they need to integrate into their work processes and which ones they can use with citizens. In this sense, a properly trained Public Administration can become a real propeller of innovation.

The fourth challenge is related to data which are the fuel of Artificial Intelligence. First of all, these must be of good quality, exempt as much as possible from biases produced by humans. Biased data can easily influence algorithms. This is why is critical to secure a right organizational methodology and create the best conditions in the contexts in which the data is produced. This is also true for data collected by IoT devices and sensors that are fragmented, heterogeneous and not highly interoperable, despite being connected to one another.

Furthermore, there is the so-called linked open data of public bodies, a real mine of information that would be very useful to generate applications of Artificial Intelligence at the service of the citizens. However, they must first be retrieved and filtered by means of semantic technologies and shared ontologies. Lastly, precisely for this kind of data, we highlight the need to ensure equal and non-discriminatory access to anyone wishing to use it.

The fifth challenge is the legal one. As always, when it comes to how to regulate the activity of the public administration, it is focused on the balance between the interests of the community and those of the individual. In the field of Artificial Intelligence is necessary to reconcile the principle of transparency of administrative acts and procedures or the protection of personal data with the right to privacy. A second issue that the public sector has to face is the relationship between the necessity of guaranteeing transparency in the context of algorithmic decision making and protecting the copyright of the creators of the algorithms. Moreover, in the event that the public administration uses programs that help the decision-making process, or programs able to take autonomous decisions, it is necessary to face the issue of accountability.

The White Paper addresses all these issues and provides some technical solutions that are also included in the European Regulation on the protection of personal data (GDPR), which will become effective in all EU countries on May 25th 2018. In general, the principles to be followed are: the transparency of algorithms and related databases, the definition of the connected responsibilities of users and the need to prevent that the use of data in the Public Administration generates a pervasive social control, in contrast with the fundamental rights of the citizen.

The sixth challenge falls directly under the responsibility of the Public Administration as it is linked to the adoption of AI technology in the public sector, for instance to improve the relationship between the State, citizens and businesses. As anticipated in the third challenge, the White Paper underlines the importance of training public employees, particularly officials and managers, on the functioning, benefits, as well as ethical and technical implications on the use of AI technologies in the public sector. The basic principle is that Artificial Intelligence should support people and help them carry out their activities, without replacing them.

This last point anticipates the theme of the seventh challenge: preventing inequalities. Indeed, Artificial Intelligence solutions must be accessible to everyone and it must be of simple and immediate use. This kind of technology can reduce social inequalities in the field of education and training, health and disability, knowledge and human rights. However, Artificial Intelligence can also increase inequalities, if the data it feeds on or the algorithms that make it up are affected by discriminatory bias. Therefore, the Public Administration must pay great attention to the development of inclusive, accessible, transparent, not discriminatory and free from bias solutions.

In order to verify the actual benefits deriving from the use of Artificial Intelligence in the Public Administration, the eighth challenge is connected to the measurement of the impact of AI technology. Once again, the challenge is presented from the point of view of both citizens and institutions. In the first case, the focus is on improving people’s quality of life and customer satisfaction; in the second case, on rendering the organizational processes more efficient and effective. In this section, emphasis is placed on the necessity to conduct both multidisciplinary quantitative and qualitative research. Indeed, the impact of technology on citizens and institutions has different facets, including economic, technical, social, cultural, psychological and anthropological factors. These variables are not always measured by the Public Administration, however, the introduction of Artificial intelligence in the public sector is such a sensitive issue that requires an attentive impact assessment.

The last challenge, the human being, is linked to the development of a discourse around Artificial Intelligence. Indeed, citizens and Institutions should be aware of the significant importance of these tools. To this end, experiments are proposed in the fields of design, arts, psychology, anthropology and sociology that can close the gap between research, industry, and society. Artificial Intelligence represents simultaneously a technological and a social innovation that can radically transform our world.

Taking this into consideration, the White Paper includes a set of recommendations drafted by the Agency for Digital Italy, the Task Force and a community of more than 500 experts and citizens. The objective is to share the recommendations with the readers in order to raise a long lasting discussion that should accompany the processes of development, modernization and improvement of the State and our society.

Eudaimonia (gr. εὐδαιμονία, lett. “being in the company of a good demon”), as illustrated by Aristotle defines human well-being as the highest virtue for a society. Eudaimonia can also mean “prosperity”, as it denotes an overall condition of well-being in which human beings perceive their benefits starting from the conscious contemplation of ethical considerations thath elp us define how we wish to live. Whether our ethical substratum is Western (Aristotelian, Kantian), Oriental (Shinto, Confucian), African (ubuntu) or attributable to any other tradition, creating autonomous and intelligent systems that explicitly respect the inalienable human rights and cultural values of users, it is possible to give priority to the improvement of human well-being as a parameter for progress in the “algorithmic age”. Recognizing the potential of a holistic approach, prosperity should in this way become more important than the pursuit of one-dimensional objectives such as increased productivity or a country’s GDP growth. Source: The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems, 2017, IEEE - Institute of Electrical and Electronics Engineers Ref. (https://standards.ieee.org/develop/indconn/ec/ead_executive_summary_v2.pdf)