ARTIFICIAL INTELLIGENCE, EUROPEAN UNION IN BALANCE BETWEEN COMMON APPROACH AND NATIONALISM

08/04/2020

By publishing the White Paper [1] in February, the European Union opened a public consultation (due to expire on May 19, 2020) about the Artificial Intelligence in order to assess the possibility of adopting a common regulatory approach to overcome nationalism and partial visions of the issue. According to the Commission, the development of common legislation on Artificial Intelligence would enable businesses and citizens of the Union to face more consciously the most current social challenges such as the fight against climate change, challenges related to sustainability and demographic change, the protection of democracies and the fight against crime. However, will these issues still be perceived in the same manner to be topical following the advent of the health emergency resulting from the spread of the COVID-19 virus?

 

Artificial Intelligence: definition and status of the EU legislation
As part of the definition of a common strategy for the digital single market, the European Union began to take a more active interest in Artificial Intelligence (AI) in 2017, when the European Council recognized the issue as urgent and rapidly evolving, inviting the European Commission to lay the foundations for a common legislation. According to the Council, it was necessary to approach AI in a uniform manner throughout the Union „while ensuring a high level of data protection, digital rights and ethical standards [2]“.
Hence, the Commission responded to the Council’s invitation and on April 25, 2018 [3] published a Communication to the other European institutions concerned in which it made public the IA strategy that would lead to an exponential increase in the Union’s resources allocated to the development of IA projects (almost 2.5 billion euro between 2018 and 2020, in addition to private investment and an economic support plan for the decade 2020-2030), also with a view to making IA-based technologies as accessible as possible to citizens and businesses. This initiative is clearly aimed at competing with other powers on the globe (USA, China, Japan and Canada) in the technology race to exploit the opportunities offered by the „home“ development of AI.
In short, the European industry “cannot miss the train”, borrowing the Commission’s words. This was reiterated by the Commission itself in its subsequent Communication of December 7, 2018 in which it set out its coordinated plan on IA [4].
Setting this apart, the Commission’s Communication of April 25, 2018 contains a (although imprecise) definition of AI, which was absent until then in the legislative texts of the Union. Scholars and technicians have provided partially different definitions over the years, but no agreement was reached on one specific definition [5]. In the Communication, the Commission defined IA as „systems that show intelligent behaviour by analysing their environment and taking actions, with a certain degree of autonomy, to achieve specific objectives [6]“. This definition was then further elaborated by the so-called Independent High Level Expert Group on Artificial Intelligence [7] which, in its report „Ethical Guidelines for Reliable AI“, specified the notion of artificial intelligence as follows: „Software systems (and possibly hardware) designed by humans that, given a complex objective, act in the physical or digital dimension by perceiving their environment through data acquisition, interpreting the structured or unstructured data collected, reasoning on knowledge or processing the information derived from this data and deciding the best actions to be taken to achieve the given objective“. To this definition it was added that „AI systems can use symbolic rules or learn a numerical model, and can also adapt their behavior by analyzing the effects that their previous actions have had on the environment“.

White Paper on IA, pillars to promote common regulation
It is clear that the adoption of a common regulation desired by the Council in 2017 on AI has been abruptly accelerated by the publication of the White Paper on AI by the Commission on last February 19th. The public consultation launched by the Commission will expire on May 19, 2020, date in which proposals and comments on how to further boost AI research and development, improve the development of AI knowledge by European small and medium-sized enterprises and provide the essential elements for a legislative framework on AI can be received by the Commission itself. In the Commission’s view, AI systems can help the Union to address current social challenges such as the fight against climate change, challenges related to sustainability and demographic change, the protection of democracies and the fight against crime. All this without neglecting respect for fundamental human rights, such as human dignity and the protection of individuals‘ privacy.
The Commission therefore intends to approach the AI issue through common legislation based on the development of the European economy by exploiting the large amount of data available through “trustworthy” AI systems that can guarantee fundamental human rights. The economic system of “excellence” that the Commission intends to develop is based on rationalization of research, on the promotion of collaboration between Member States (as well as between the public and private sectors), increasing investment in the development and dissemination of AI.
In order to make IA systems “reliable”, the Commission has confirmed the Group of Experts’ view that any development of “reliable” IA cannot be without ensuring the following seven key pillars:
(i) human contribution and supervision; (ii) technical soundness and security; (iii) privacy and data governance; (iv) transparency; (v) diversity, non-discrimination and fairness; (vi) social and environmental well-being; (vii) accountability.
This is obviously to prevent the development of AI-based systems from having distorting effects. Indeed, the Commission is well aware that AI systems, through the automation of activities that were previously the exclusive competence of humans, may threaten the privacy of citizens, infringe their right of expression, the principle of non-discrimination and the self-realization of individuals (just to mention some of the risks).
However, the fact that AI systems are able to process an immense amount of data through automated algorithmic processes (predetermined by humans) can be a valuable help to counter the spread of the COVID-19 virus in the current health emergency situation, leading individual Member States to „run alone“ on the issue of the medical emergency.

Medical AI systems – COVID – 19 case
It is no surprise that, in the medical field, the most advanced diagnostic systems are able to process by themselves and in real time a myriad of data from the Internet, public administrations and other networked health facilities on their own. This has been happening for a long time in this area, so much so that it is common to talk about the Internet of Medical Things [8]. Obviously there are many ethical problems underlying the use of AI in the medical field both as reliability of the data processed (acquiring patient data from the Internet or other databases if not sufficiently validated could lead to incorrect diagnoses) and as objectives pursued (AI could pursue unethical objectives such as driving towards medical practices that meet administrative objectives but not the real quality of care) [9].
The White Paper sets the whole health care sector as an example for a high-risk area that should be taken into account in any future regulation.
In the situation now described, the health emergency resulting from the COVID – 19 infection, which – with a shocking effect – brought to light two crucial issues: the first concerns the new balance between the medical field and fundamental rights in the emergency health context, while the second concerns the real possibility of continuing, and how, with collaboration between Member States in the European context.
In clear contrast to the indications of principle set out in the White Paper, in fact, in the current health emergency to combat COVID-19 there is a progressive complication of the ways of balancing the pillars of ethics and the protection of fundamental human rights with the objectives of protecting the right to health, precisely in that high-risk field identified in the White Paper. This could mean, on the one hand, that the path of regulation based on ethical principles indicated by the Commission will be subject to harsh criticism for failing to take account of what was already happening in the medical field in relation to AI. In particular, the Commission makes no mention of the necessary balance between fundamental human rights (such as privacy) and the necessary applicability of AI systems, which necessarily have to sacrifice citizens‘ privacy in order to protect public health. On the other hand, we are already seeing clear positions taken by individual Member States in favor of the applicability of AI in the medical field, with the consequent sacrifice of fundamental human rights as set out in the White Paper, such as privacy.
To confirm this, see the discussion underway in Italy where, among other mandatory measures to be taken, the possibility of using a smartphone application capable of tracking citizens‘ movements to stop the COVID-19 epidemic is being considered. Obviously, the privacy’s breach connected to the use of the aforementioned application is clear. Even more clear is the use of AI capable of processing the data received by providing a diagnosis in a short time to contain and fight the pandemic without citizens being able (perhaps) to oppose the tracking of their personal and medical data. In addition to the tracking via smartphone application, the Ministry of Technological Innovation and Digitization, the Ministry of Economic Development and the Ministry of University and Research have published an invitation to research centers and innovative companies to provide a contribution in the field of devices for prevention, diagnostics and monitoring to contain and combat the spread of the COVID-19 virus (including the possible use of a monitoring smartphone application) [10]. On this point, the approach of Antonello Soro, President of the Guarantor for the protection of personal data, appears wavering. On the one hand, in the first days of the emergency, he was in favor of using the app in question if such a system of data collection and processing, even if invasive, is in any case aimed at the general interest of health protection [11]. On the other hand, recently, the Guarantor has expressed an opinion that would appear to be different, arguing that the use of apps of this type can only take place on a voluntary basis. Now, it would seem appropriate to approach the issue from a holistic perspective, asking whether or not the health emergency justifies the compression of fundamental rights. If this is the case (and it seems that Italian constitutionalists have expressed themselves in this sense), the compression of the right to privacy is no different from that of the right to freedom of movement, of association, and of work that citizens have already been suffering for several weeks, for the – at least allegedly – superior good of public health. It is obvious that the compression can only be justified to the extent that the health emergency actually exists, and that the two opposing demands must be constantly balanced, also from an evolutionary point of view. When the threat is greatest, more invasive measures can be justified; when the threat is least, these measures can become illegitimate. It is equally obvious that it will be necessary to ensure adequate control of the choices that in these days are made by governments, regional governors, other administrative authorities, with little or no involvement of Parliament and in the paralysis of the judicial system. Both these aspects must be rapidly restored, on pain of upsetting the fundamental characteristics of our society. It will be interesting to see what decisions will be taken by Italy and the other EU Member States. In fact, it is to be expected that the latter will undertake their own paths to implement AI systems in the medical field to counter the spread of COVID-19.
And here comes to light the second issue mentioned above, namely the disruptive effect that the COVID – 19 emergency is having on all the structures and principles of the European Union. Member States soon suspended some of these principles, first of all that of free movement of people and goods, blocking borders, deliveries of goods considered essential, and sometimes seizing such goods in transit to supply their own facilities. Also in the field of IA and apps for health control it is foreseeable that each Member State will move autonomously, with good peace of mind of the initiatives undertaken by the Community bodies, and highlighting how the whole system must now be rethought.
A fundamental question therefore remains to be asked: will the fundamental pillars – also in terms of AI – developed by the European Union stand up to the challenge of the health emergency that has upset Europe?

Footnotes:
[1] Full text available at https://ec.europa.eu/info/sites/info/files/commission-white-paper-artificial-intelligence-feb2020_en.pdf.
[2] Council document EUCO 14/17 CO EUR 17 CONCL 5 of 19 October 2017, full text available at https://www.consilium.europa.eu/media/21608/19-euco-final-conclusions-it.pdf
[3] COM (2018) 237 of 25 April 2018, Communication from the European Commission „Artificial intelligence for Europe“ full text available at https://ec.europa.eu/transparency/regdoc/rep/1/2018/EN/COM-2018-237-F1-EN-MAIN-PART-1.PDF.
[4] COM (2018) 795 final December 7, 2018, Communication for Coordinated Plan on Artificial Intelligence available at the link https://ec.europa.eu/knowledge4policy/publication/coordinated-plan-artificial-intelligence-com2018-795-final_en.
[5] For further details see LAVAGNINI, Intelligenza artificiale e proprietà intellettuale: proteggibilità delle opere e titolarità dei diritti in Il diritto d’autore 2018, fasc. 3, 360 ss
[6] The Commission also specified that AI-based systems can „only consist of software that acts in the virtual world (e.g. voice assistants, image analysis software, search engines, voice and facial recognition systems), or incorporate AI into hardware devices (e.g. advanced robots, self-driven cars, drones or Internet of things applications)“.
[7] The High Level Expert Group on Artificial Intelligence is independent and was set up by the Commission in June 2018 and made its opinion public in April 2019, available at https://ec.europa.eu/digital-single-market/en/news/ethics-guidelines-trustworthy-ai.
[8] On this point see https://www.europeanpharmaceuticalreview.com/article/47692/imot-healthcare/.
[9] For an examination of the advantages and risks see Musacchio, Guaita, Ozzello, Pellegrini, Ponzani, Zilich, A. De Micheli, Intelligenza Artificiale e Big Data in ambito medico: prospettive, opportunità, criticità, AMD Journal, ISSN 2036-363X (print) vol. 21-3.
[10] The project called “Innova per l’Italia” is published on the website of the Ministry for Technological Innovation and Digitization at https://innovazione.gov.it/innova-per-l-Italia-la-tecnologia-e-l-innovazione-in-campo-contro-l-emergenza-covid-19/.
[11] On this point see www.garanteprivacy.it/web/guest/home/docweb/-/docweb-display/docweb/9298389.

Alessandro Bura, Tankred Thiem