Challenge 8: Measuring the impact

“Change has never happened this fast before, and it will never be this slow again” [1]. The introduction of new Artificial Intelligence (AI) technologies in the society and, in particular, in the Public Administration, brings with it the natural desire to measure and understand its social impacts, risks and opportunities.

It is now essential to measure the impact of public policies, both in terms of the user, i.e. the citizen, and the PA. Regarding the first point, it is necessary to think in terms of improving the quality of life of people, but also the conditions of use of what is offered to them.

The measurement of the impact in using Artificial Intelligence solutions in the PA contemplates the use of qualitative and quantitative indicators. For example, the methods for measuring customer satisfaction (e.g. social impact, well-being of citizens, accessibility and usability of the tools) or related to the optimization of organizational processes in terms of efficiency and effectiveness

Many quantitative models subdivide workers by their employment and try to hypothesize which professions will be replaced by technologies, in other words these models base their operation by considering jobs and employment as a unit of analysis [2]. However, technology often does not completely replace a professional figure but replaces only some specific activities. Workers who previously held a particular task are therefore addressed and reassigned to complementary activities that use the new technologies. Over time, technology leads to a complete rethinking of organizational processes and objectives.

Given the complexity of the phenomenon to be analysed, the impact must be measured necessarily taking into account a multidisciplinary approach, which allows defining the impact also from an anthropological, psychological and sociological point of view, as well as from a technological and econometric point of view.

For this reason it is necessary to identify new sets of indicators that can better adopt this multidisciplinarity, in synergy with the indicators existing today.

In any case, it is necessary to keep in mind that the methods adopted to measure the impact can promote a better understanding of the services by the users and encourage the transition to new governance models [3].

Mapping the needs and defining the impact objectives with all the actors involved, collecting real-time information on how all the nodes of a network interact, are the first essential steps in understanding and defining correct policy assessments.

Unfortunately, these assessments are not updated so frequently as they should, due to financial limitations or unavailability of competent assessors. As policy assessment is commonly based on data, AI could enable a faster and more accurate analysis.

A greater balance in the analysis of opportunities and risks could be an important factor to increase awareness of the real impact [4]. that AI can have in our daily lives.

Finally, measuring the impact of these technologies is useful in terms of designing and developing AI, in order to guarantee reliability and transparency as well as to reduce the risk of errors, also as regards the Public Administration [5]. Analysing its operation allows us to determine valid models for an ethical and responsible use of AI.

Measurement tools must be strict to determine the effective social impacts of AI and to define how these technologies can actually influence our lives.

Footnotes

[1]Justin Trudeau, Canadian premier, in his speech at the World Economic Forum Annual Meeting 2018 Ref. <https://www.weforum.org/agenda/2018/01/pm-keynote-remarks-for-world-economic-forum-2018>`__ (Consulted in February 2018).
[2]Ref. Challenge “Skills” and Challenge “Reducing inequalities”
[3]Ref. Challenge “Accompanying transformation”
[4]European Commission - DG for Research and Innovation, Directorate A -Policy Development and Coordination, Unit A.6 - Open Data, Policy and Science Cloud; “Vision and Trends of Social Innovation for Europe”, 2017
[5]Ref. https://www.researchgate.net/publication/23542471_Spatial_diversity_in_local_government_re venue_effort_under_decentralization_A_neural‐network_approach.