Responsible AI and welfare

Responsible AI and welfare

Person holding green piece of paper with the letters A.I. written on. Photo.

The focus of this EiT village is to establish new knowledge about the use of Artificial intelligence (AI) applications in public administration in Norway focusing on the users´ need to understand the role of AI applications in welfare delivery processes, their key characteristics and the types of data they use. This will lead the way in developing human-friendly and trustworthy solutions.

The use of AI applications for the delivery of public services creates the need for a more human-centred perspective. On one hand, the citizens should have an understandable, transparent, and trustworthy experience interacting with AI applications and on the other hand, the public service employees involved need to make sense of the logic of the AI systems that are used to support their everyday work for public service delivery. European Regulation (GDPR) and the Norwegian Public Administration Act stipulate that the logic involved in decisions must be communicated (i.e. the grounds for decisions must be given, facts used for decisions must be provided and algorithmic support must be explained).

The village will attract and motivate students from different disciples who will understand the challenges from different perspectives (for instance: technical, legal, social, business, psychological and design perspectives), do research, propose their own ideas, and project assignment within the thematic area, ending up delivering the required outcomes. This can foster student creativity and a strong sense of ownership of the conceived project. Student teams will be able to focus for example on empirical studies, creation of personas and prototypes, development of design principles or other.  
 

Two people working in teams in front of laptop on a table. Photo.Relevant competency

The goal is to create teams of students with different profiles, who will be able to address challenges through collaborative thinking. Everyone interested in helping to tackle challenges for the responsible use of AI in public sector, interested in taking part in a process that involves a collaboration with citizens and public institutions.
The breadth and scope this village are suitable for participants of a wide range of backgrounds.  Computer Science and IT, Artificial Intelligence, Software Engineering, Psychology, Health, Law and Social Sciences are particularly relevant towards the direction of the village, but other disciplines are also welcome.

About the village

Artificial intelligence (AI) applications nowadays are widely used and can increase the efficiency and quality of processes across sectors including public services. However, their inner workings are not easy to be understood and thus making it hard to explain how they transform data inputs to the outputs the AI application may provide. This significant challenge is known as the “black box” problem which can hinder the involvement of humans in shaping, operating, and monitoring the use of AI in welfare service delivery. The focus of this village is to investigate the current challenges related to the AI applications in public services aiming to help users (case handlers in welfare organizations and citizens) to understand, trust and effectively manage new AI applications. This will lead the way in developing human-friendly and trustworthy AI.

Challenges related to the above-described topic for this village involve the two key stakeholder groups which have different needs:

1) Case handlers at the operational level need to comprehend the decision-taking logic. Furthermore, they need to be aware of the limits of the particular algorithmic support. It is important to make clear for case handlers what are the key attributes of the systems used. For instance, is the system static or keeps learning using newly accumulated information?

2) Citizens need to be aware of the methods used for decisions that affect them. Like the case handlers, citizens need to understand some key attributes of the systems in use. Furthermore, citizens need a level of insight to the types of data used especially when there is a risk that data may be incomplete or unrepresentative of those who may be subsequently affected.

Possible research question:

  • How to inform non-experts about the role and characteristics of AI used for welfare service delivery facilitating meaningful human control?

Example of challenges:

  • Elicit the needs of different citizen groups and case handlers
  • Create solution ideas or prototypes that visualize information in an understandable way
  • Evaluate different alternative visualizations and their impact in decision taking or in citizens´ acceptance
  • Increase the efficiency of decision processes without compromising transparency
  • Respect the law and build trust to the systems and case handlers.

The village is collaborating with the group of researchers from University of Agder (UiA) and NTNU. The purpose of the partnership is to the provide village participants with information, facilitate a forum for discussion and evaluation of proposed actions to specific problems related to the responsible use of AI for the delivery of public welfare services.

 

Facts

  • Course Code: TDT4850
  • ​​​​​​​Type: Intensive
  • Teaching method: Classroom teaching
  • Language: English
  • Leader: Sofia Papavlasopoulou
  • Contact: spapav@ntnu.no ​​​​​​​
  • Semester: Spring 2022
  • Host faculty: IE

How do I register for EiT?

Important information about EiT:

Important information about EiT:

  • The focus on teamwork skills and group processes is the unique feature of Experts in Teamwork (EiT).
  • EiT's teaching methods depend on the contribution and presence of every participant throughout the semester. For this reason, attendance is compulsory on every village day.
  • In contrast to many courses, the first few days are especially important in EiT. During this period, the team members get to know each other and discuss what each individual can contribute. You will also draw up the compulsory cooperation agreement and start preparing a shared research question.
  • For additional information about Experts in Teamwork, see the page for students.