Artificial intelligence and public regulation, with special attention to the local world: sixth Living Lab of Good Governance and Quality of Regulation

On 18 November 2024, the sixth Living Lab on Good Governance and Regulatory Quality took place at the headquarters of the Bosch i Gimpera Foundation of the University of Barcelona (about the previous ones, see here).

As is well known, a living lab is a system for innovating on social challenges in a collaborative way. The main differences between the Living Lab model and other traditional participatory models are that it provides more creative contexts, with horizontality and empathy; more representative, as the participating community (which goes beyond the purely academic) is articulated according to the challenge to be solved; and more innovative, as it seeks experimentation and co-creation with longer lines of maturity. Moreover, this type of Living Lab can become a hub for sharing information and experiences. Increasingly, universities are using this methodology in collaboration with the public, private and third sectors.

On this occasion, the Living Lab focused on artificial intelligence and public regulation, with special emphasis on the local level, bearing in mind the new EU Regulation on this subject, published in the BOE on 12 July.

The programme was divided into three blocks with several moments for the exchange of opinions among the dozens of attendees, both in person and online:

These blocks started with a presentation and were developed by various professors, who introduced the main issues and problems and led to discussions with members of the public and private sector, moderated by one of the co-directors and the coordinator of the Yearbook:

Beyond the short review below, the full video of the activity is available here.

Professor Cerrillo introduced the relationship between IA, local public policies and existing regulation, including the European Regulation, which foresees the future elaboration of Codes of Practice, and there is already a draft of one of them referring to general IA, as highlighted in the talks. As is known, Codes of Practice are foreseen in Art. 56 of the Regulation and are being developed by the IA Office and a wide range of stakeholders. Reference was also made to the Tarragona Digital Administration Ordinance or the Barcelona City Council Protocol on IA. On the other hand, reference was made to the need to individualise the development of AI locally through ethical codes dealing, for example, with the use of the well-known ChatGPT. In the ensuing discussion, issues such as, for example, the possible usefulness of a Guide for the use of AI in the local world, along the lines of the one that exists in the United Kingdom for the judiciary, were highlighted.

In the second block of conversation, Profs. Miranzo and Galindo introduced ideas on AI and its impact on the local world in the areas of recruitment and public employment. The conversation highlighted how the wave of retirements expected in the public sector has in many cases not changed the way in which future public employees are selected, of which, until now, those at levels C1 and C2 have formed the backbone of the Public Administrations. The change of scenario brought about by AI requires a change in both aspects, with an increase in the intake of new profiles of public employees and continuous training to improve AI literacy (art. 4 of the European Regulation).

On the other hand, the purchase of AI systems in the private sector entails the risks of public dependence on the private sector and the displacement to the private sector of decisions that should be taken by the authorities, as well as possible cases of opacity due to the limits of AI transparency, for example because of intellectual property rights.

In any case, the importance of the assistance that can be provided to small municipalities in relation to IA by supra-municipal bodies was highlighted. In this sense, reference was made to art. 36.1 g of the LBRL, which refers to the provincial competences regarding ‘the provision of e-government services and centralised contracting in municipalities with a population of less than 20,000 inhabitants’, and the need to reflect on this issue.

Likewise, the problem that arises if an Administration intends to use data obtained for a purpose of general interest in relation to another, such as the development of a proactive activity through AI (for example, the granting of ex officio aid), was also highlighted, in relation to the need for consent for the use of such data.

In the third and last block, Ms. Ascen Moro presented the municipal experiences in the use of AI, especially in Sant Feliu del Llobregat City Council, where it is used to predict social vulnerability or the needs of the elderly. Reference was also made to the Atenea programme as an element in the fight against the digital divide. Prof. Martínez and Dr. in Philosophy and lawyer Javier de la Cueva closed the introduction to this block with references to the patrimonial responsibility of the Administration and the judicial control of the use of AI by the Public Administrations.

Prof. Martínez referred to the positive aspects of the system of objective financial liability theoretically existing in Spain, raising problematic issues linked to damages caused by bias, data leaks, information provided by bots or material damages caused by material or technical activities (drones, health field…). He also referred, among other issues, to the role of the progress clause of art. 34 LRJSP or to the more general question of whether the fear of liability could slow down the development of AI in the local public sector, as well as the interest of considering aspects of liability linked to the intervention of a contractor.

Javier de la Cueva discussed various issues related to the civilisational change that AI represents, the problems in the control of power in the 21st century in the face of large technological companies, which represent a change of paradigm with respect to the Law that emerged from the liberal revolutions of the 18th and 19th centuries, for which our Constitutions, with their traditional declarations of rights, are not prepared. In the ensuing conversation, various topics linked to issues of responsibility and judicial control arose, with references to the well-known BOSCO case, which has already given rise to two rulings, in 2021 and 2024.

The activity closed with the forecast of the future publication of the Yearbook of Good Governance and Regulatory Quality, which will gather the contributions of the living lab and will be published before the summer of 2025.

Finally, remember that the full video of the activity is available below: