Artificial intelligence in public administration. How to regulate and not discriminate?
Closing the first day of our conference on artificial intelligence and human rights we had Lluís Torrens, from the Innovation Department in the Social Rights Area of Barcelona City Council, and Cyntia Menezes, a lawyer specialized in digital law at Font Advocats. The session was moderated by Karma Peiró, a journalist specialized in technology.
Peiró began the meeting with some thought-provoking questions about AI in the public sector: What policies and ethics do public administrations apply with regard to the use of the algorithms? Will there be seals of quality so that citizens can have confidence in these algorithms? How do we know if an automated decision-making algorithm is being applied to us?
Lluís Torrens highlighted how Barcelona City Council is using technology to assist people in vulnerable situations, particularly in the field of social rights. “Technology allows us to capture and analyse large volumes of data and with Artificial Intelligence we get more than just statistics”. Torrens explained in detail the five lines of action that you can see in the video. Broadly speaking, the actions are categorised as follows:
- Big Data Social: “we take data from our users and put it in a large repository to operate the action plans”
- Welfare simulator: “we use them to make it easier for people to know anonymously, and without prejudice, whether they are entitled to state aid”
- Apps: “the way to establish new communication channels between users and professionals”
- Chatbots: “designed to assist people in their consultations and to collectivize information of general interest to avoid queues or have people answering the same questions”.
- Collective intelligence: “we prefer this term because everything behind artificial intelligence has been created with years of experience of professionals”
It was precisely on this last term of collective intelligence that the rest of Torrens’ presentation revolved, and we encourage you to listen to it in full.
Cyntia Menezes contributed key reflections on the regulations that governments and public administrations create around technology and automated decisions. Menezes began by warning that this view would perhaps frustrate the expectations of Peiró, Torrens and the people present regarding the responses that AI laws can provide us. “There is still little regulation on Artificial Intelligence and automated decisions. There is a lot of soft law (recommendations, guidelines, studies, ethical codes…) but it has no legal applicability,” laments Menezes.
Of all the regulations at a global, European and Spanish level that Menezes reviewed, the lawyer wanted to highlight one concept above all else. The design of AI systems that incorporate “ethical reasoning“. “We talk a lot about privacy but we have to be careful because when we talk about Artificial Intelligence in legal universes this concept goes far beyond just data protection,” warns Menezes. “Artificial Intelligence does not work without data, it is its fuel!
Menezes reviewed several fields in which the use of Artificial Intelligence can have important effects, noting that its misuse can constitute a “global problem” and prevent countries that are still developing from achieving technological autonomy. If you want to hear what Menezes thinks are the crucial aspects to take into account when governments use AI, and what challenges lay ahead for civil society, do not miss the full video!
Here are some links to resources commented on during the meeting:
- Informe: “Intel·ligència Artificial. Decisions Automatitzades a Catalunya”. (2020)
- Automating Society: Taking Stock of Automated Decision-Making in the EU
- “How to (partially) evaluate automated decision systems”. Data Justice Project. Cardiff University
- Oxford Artificial intelligence in hiring: Assessing impacts on equality
- AI WHITE PAPER On Artificial Intelligence – A European approach to excellence and trust
- Renta mínima y ‘big data’: a first step towards the public policy revolution