Design for Responsible AI and welfare: 2022 EiT Village

14.01.2022 – 28.01.2022, NTNU Trondheim, Norway
Course teacher (village supervisor): Sofia Papavlasopoulou
Course code: TDT4850

The objective of this Design activity was to establish new knowledge about the use of AI applications in public administration in Norway, focusing on the users´ need to understand the role of AI in welfare delivery processes, their key characteristics and the types of data they use, leading the way in developing human-friendly and trustworthy solutions. Participants worked alongside researchers from the AI4users project from both University of Agder (UiA) and NTNU. The purpose of this partnership was to provide participants with relevant insights, facilitate meaningful discussion, and evaluate proposed measures addressing specific challenges related to the responsible use of AI in delivering welfare services.

EiT is an obligatory course for study programmes at the master’s level at NTNU. In this course, students from different study programs come together to solve real challenges through collaboration, reflection, and shared responsibility. It works by placing diverse groups in interdisciplinary villages, where they explore problems, develop ideas, and practice communication, trust building, and constructive conflict resolution. This takes place through written and oral reflections, and structured teamwork exercises. Through this process students discover how their individual strengths become far more powerful when combined with the perspectives of others, and they learn what effective teamwork truly means in practice. The teaching takes place in courses (villages) which each have a different focus and challenges.
Read more about EiT on NTNU´s website.

The teams had to pick one of three suggested cases, and design how they would present the described AI model to the citizens. The three cases were:

  1. AI model suggesting what kind of support is better for an unemployed person (e.g. retraining, coaching for job interviews, only unemployment stipend).​
  2. AI model suggesting high-risk cases of student loan holders that need to be checked to minimise fraud.​
  3. AI model suggesting the expected total duration of a sick-leave (is the end closed or not) so appropriate support can be given (e.g. preparation for coming back to work).​
    All cases have been discussed in Norway (but none were launched at that time).

Timeline