Design of trustworthy and inclusive AI services in the public sector

Research output: Contribution to conferencePosterScientific

Abstract

Algorithmic decision-making and big data systems are increasingly being used to provide innovative and essential services in the public sector. A recent report documented 686 use cases of AI-based public services among most European countries [1]. How to design AI-based services to ensure they are trustworthy, inclusive and empowering for civil society? Moreover, what are the perspectives on this of practitioners who design and develop AI-based public services?
The distinct regulatory demands and responsibilities of public sector services require crucial consideration of inclusiveness. The need for the trustworthiness of AI-based public services is highlighted in several European reports [2, 3]. Previous research shows that for citizens to trust new AI-enabled services they require greater agency [4]. However, I found a sparse body of work specifically addressing inclusion and civic empowerment in AI-based services [5, 6, 7].
I conducted in-depth interviews with ten practitioners who are responsible for managing, developing or designing AI-enabled public services across three big public organizations in Finland. I examined their attitudes, practices and challenges of implementing trustworthy, transparent, and inclusive AI services that can empower civil society. As the next step, I plan to perform a co-design of a public AI service with representatives of civil society and public sector actors.
The current results show that the discussion on inclusion and civic empowerment is just in its beginning in the public sector. Practitioners perceive the concept of inclusion as devising accessible public services for all members of society. Civic empowerment was understood as 1) institutional transparency, 2) civic participation in shaping the services, and 3) easing the use of the services for their users. The research suggests two distinct socio-cultural constructs that may hinder how civic empowerment is manifested in such services: expert and risk-averse cultures.
This work in progress might become a vital contribution to Nordic public sector actors, helping in achieving trustworthy, inclusive and empowering AI-based public services.
Original languageEnglish
Publication statusPublished - 2 Oct 2023
MoE publication typeNot Eligible
EventNordic AI Young Researchers Symposium - National Museum of Denmark, Copenhagen, Copenhagen, Denmark
Duration: 2 Oct 20233 Oct 2023
Conference number: 3
https://2023.nordicaimeet.com/

Conference

ConferenceNordic AI Young Researchers Symposium
Abbreviated titleNAIM
Country/TerritoryDenmark
CityCopenhagen
Period02/10/202303/10/2023
Internet address

Fingerprint

Dive into the research topics of 'Design of trustworthy and inclusive AI services in the public sector'. Together they form a unique fingerprint.

Cite this