Professor Slavkovik: AI is Key to a Better Future

An artificial intelligence (AI) revolution, spearheaded by tools like ChatGPT, suddenly  hit the world a few years ago. While some are concerned about jobs disappearing and students using AI to cheat on their exams, one of the leading AI researchers at UiB, professor Marija Slavkovik, head of the Department of Information Science and Media Studies at UiB, believes there is a cost to not leaning into AI that we should not overlook.

Slavkovik is active in the AI subdisciplines of multi-agent systems, machine ethics, and computational social choice. She believes there are mostly benefits to the technological shift we are amid.

«Europe has less and less people, but we want those people to enjoy the current lifestyle or better and have freedoms. There is a cost to not automating. That means we are going to have worse and worse experiences in our society with the public services – simply because we are running out of people,» she says.

Changing Society for the Better

«If we want to have services in society, work a reasonable number of hours and have the affordance to pursue our passions, how do we get there? The same way we have gotten there to begin with. By automating things,» she says, and adds:

«But of course, we cannot have blind automation without considering individual and societal short term costs.»

«What about productivity? What about jobs disappearing?»

«I think this is a very short-sighted view of things. The point is to create more productivity by using AI. There is a systematic way to go about these things. There will still be a need for jobs, but what we do in those jobs is going to be a little bit different,» she says.

Technology has usually changed society for the better, according to Slavkovik.

«You are not going to convince me that we lived better in the past. Some people did, but not most. People have an optimistic view of the past and a pessimistic view of the future,» she says. 

Trust and Algorithmic Decision-making

At the moment there is a need for a lot of research on artificial intelligence. Slavkovik wants to be a part of the solution. She is hoping to get an ERC Synergy Grant to study algorithmic decision-making in public institutions. She is now in the last round of interviews.

«Specifically, we want to look at the trust relationships between citizens and public institutions, and how they are disrupted when AI gets involved,” Slavkovik explains.

Algorithmic decision-making (ADM) is when someone uses computation to make or aid decisions about citizens. Some countries use them in hiring, some for identity checks of migrants, and some for assessing the adequacy of financial aid for paying electricity bills. Slavkovik says that from about 2020 there has been a wide use of this in public institutions.

«ADM is not just AI. It uses AI, but it is a little bit more than that. It is basically something that uses a statistical model trained by AI to generate information about behavior. It is a very large data analysis, and the analysis is used algorithmically to make assessments. Ideally, then there will be a public servant who looks at this and says ok,» Slavkovik explains.

The Right Amount of Trust

A machine can make mistakes or unethical decisions. So why do public institutions want to use algorithmic decision-making?

«It makes the work of public administration a lot more efficient, but it is problematic because public administration has more power than private companies. Let’s say your grocery store uses it to predict whether you are going to shoplift or not. You can always go to a different store. If Nav uses it to predict whether you are going to be unemployed for a long time or not, you cannot use a different Nav,» Slavkovik says.

She is interested in questions like: How do we get the citizens to place the right amount of trust in this? How do we use policy to empower the citizen to be in control over their process and their relationship with the public institutions?

«We need to increase the amount of ADM in public service so we can have functional public service that serves the citizens, but we also want the citizens to be empowered, to make the most out of it, to know when to, and have the means to object to and contest the decisions of public services,» Slavkovik says.

«I don’t know what the overall solution can be, but we can give people a starting point. Making people aware of what they should be paying attention to, and giving an understanding of what this means is a good step forward.»

The first article of the Maggetti, Lahusen, and Slavkovik team calling for understanding trust towards better automation of decisions came out on the 5th of September. It can be accessed here.