Socially Empowered AI Research – SEAIR expands the AI research horizon
”Since so much of AI research is driven by commercial forces, societal impact often becomes an afterthought, rather than something that has to be studied on a fundamental level”
Main content
Factory automation reshaped the workplace in the 1960's. Many jobs were lost, and new products, services and commerce were gained. Because one was not going to “get tomorrow’s jobs with yesterday’s skills”, this process profoundly changed education and societal values as well. Now, the same challenge is brought to us by artificial intelligence (AI). What are we going to do about it?
Marija Slavkovik heads The Department of Information Science and Media Studies and is leading the initiative for the new AI center SEAIR.
”Since so much of AI research is driven by commercial forces, societal impact often becomes an afterthought, rather than a something that has to be studied on a fundamental level,” says Slavkovik.
Read: New AI breakthroughs show that bigger isn’t always better
Applying and refining already developed AI solutions to fit Norwegian society is an important path. But we must also realise that progress in engineering and business is always driven by progress in research. What we urgently need are design and development of completely new algorithm and knowledge representation methods that can pack societal values as data structures.
Equally important are regulations that ensure that the privileges and benefits of AI are justly acquired, distributed and enjoyed. We need to generate more knowledge: we need more AI researchers who will continue to generate more knowledge for the decades to come.
AI research can and should be better
“Bergen is already a hub for AI research. It will be great to use all this brain power across departments and disciplines, the experience we gathered in being on the absolute front of AI and AI Ethics education in Norway and all of our international collaborations and standing in the AI research field.”
Read: (NRK) Norges forstekull med KI studenter er ferdig utdannet fra Universitetet i Bergen
Large and ambitious goals
While progress in AI has been fantastic, the fundamental approaches and methodologies are the same as they were 70 years ago when electronic computing begun. The team behind the Socially Empowered AI Research centre proposal is convinced that we can and should do better.
Norway is already adopting algorithmic decision making (ADM) in many areas, such as medical diagnostics and in detecting tax and insurance fraud. This increases efficiency, but ADM – like all computers – only processes facts without its societal context. Humans, on the other hand, include moral values, cultural sensitivity and the well-being of others when making decisions. Automating decision making without these considerations may have unintended consequences in the long run.
SEAIR will examine the ADM information-processing pipeline, and work to address each element of the timeline:
- The algorithms that train prediction models
- Uncertainty and causality
- Artificial agents that embody ADM and interact with beneficiaries
- Digital twins for testing, integration and monitoring
- Value alignment of algorithms, models and agents
In the SEAIR center we will expand the horizon and establish a focal point of AI research. In this development of automation and technology the most important resource is people. We also need to pay attention to Norway being competitive internationally in attracting the people who want to be trained in AI research and who will be new hires in the Norwegian companies.
“We want SEAIR to be an attractor of the best students on all levels that want to bring AI to the next level,» says Inge Jonassen, the head of the department of Informatics which hosts the SEAIR center.
The aims for SEAIR are set high - Not only do the SEAIR researchers want to advance AI research, but also do it by a very ambitious use of interdisciplinary approaches, and cross sector approaches.
However, the centre leadership has full confidence in their goals. As well as Marija Slavkovik, the centre has vice leaders Fedor Fomin and Pinar Heggernes, as well as a fierce group of interdisciplinary young researchers. Together, they feel ready to take on the big unanswered questions of AI research.