0
Jump to main content

Strengthens responsible use of artificial intelligence in public services

The Research Council of Norway has granted NOK 15.9 million for a research project on the responsible use of artificial intelligence in public administration. 

Illustrasjonsfoto
RESPONSIBLE USE: An international team of scientists headed by UiA are focusing on how to make artificial intelligence an understandable tool in public administration and service in Norway. (Ill. photo: Colourbox)

The money is awarded to the four-year research project “Responsible Use of Artificial Intelligence through Design for Accountability and Intelligibility”, abbreviated to AI4Users. The funding comes from the Research Council 's programme IKTPLUSS for research activities within the thematic priority Ubiquitous Data and Services that move the research front, promote essential competence-building and create new opportunities. The project is led by Associate Professor Polyxeni Vasilakopoulou (picture) who participates to the Centre for Digital Transformation (CeDiT) at the Faculty of Social Sciences at the University of Agder.

Polyxeni Vasilakopoulou, foto

Polyxeni Vasilakopoulou

Responsible use of artificial intelligence

The research work at UiA stems in part from the Norwegian government's stated goal of promoting responsible use of artificial intelligence in public administration. The desire is to take the lead in the development of humane and reliable solutions. 

“The AI4Users project takes a human-centred perspective for the development of tools and design principles to help users trust AI solutions. The novelty of AI4Users is that it targets specifically non-experts extending the reach of research beyond AI experts and data scientists”, says Polyxeni Vasilakopoulou.

Artificial intelligence (AI) refers to machines performing the cognitive functions typically associated with humans, including perceiving, reasoning, and learning. It is not a specific technology but an umbrella term that includes a multitude of different technologies. For instance, AI technologies include machine learning algorithms that process data to make predictions or decisions without being explicitly programmed to do so. 

The project is performed by a consortium of collaborating partners from academia: The University of Agder (leader), the University of Oslo and Norwegian University of Science and Technology (NTNU). Furthermore, the project includes international researchers from Bentley University in the USA, Vrije University Amsterdam in the Netherlands, Exeter University in England, and the University of Gothenburg in Sweden. 

"The collaboration with international experts will offer possibilities for leveraging experiences from other countries, making comparisons and better disseminating results" Vasilakopoulou notes.

Different users have different needs

There is growing concern about the lack of transparency in AI solutions. This is frequently called “the black box problem”. The responsible use of AI can be achieved if users understand the information that such solutions possess, their capabilities and limitations.

Different users have different needs, for instance: 

  • citizens need to be aware of the algorithmic support used for decisions that affect them and the types of data used; they need to trust that data are not incomplete or unrepresentative.
  • public service employees at the operational level need to be aware of the limits of the AI solutions and of ways for handling unexpected situations. 
  • policy makers and managers have a special interest on the temporal dimension, they need to be able to monitor the evolution of AI performance over time ensuring fairness and regulatory compliance

“AI4Users is a good example of sociotechnical research for the ongoing digital transformation of our societies, this is at the core of CeDIT´s aims​” Vasilakopoulou says.

Research that emphasizes collaboration and is evaluation driven

To ensure that the research is relevant, the project will collaborate with NAV linking to NAV´s ongoing work. NAV today explores artificial intelligence to deliver better and more efficient services in a responsible way. The key to this is the establishment of NAV's AI-lab. The project is clearly aimed at users, and collaboration with NAV is central to this.

"The project will be carried out in the period November 2020 – October 2024 and will follow an Action Design Research (ADR) approach. ADR is an iterative and adaptive research method closely linked to practice. The method emphasizes collaboration and is evaluation driven" says Polyxeni Vasilakopoulou.

Illustrasjonsfoto

 About the Digital Transformation Centre (CeDiT):

  • Centre for digital transformation - CeDiT - is a social science centre at the University of Agder that researches the effects of digital transformation on societies, organizations, and individuals now and in the future.
  • The aims of CeDiT are to conduct theoretically based and empirically informed research, influence the academic discourse by publishing in internationally renowned journals, create an interdisciplinary arena for exploring the properties and consequences of digital transformation, and actively seek collaboration with others.