Artificial Intelligence, AI
The field of artificial intelligence is growing rapidly. Much progress is driven by large companies and technical institutions. The proposed solution to this problem is to put the human at the center stage of the equation by reconnecting to the Scandinavian tradition of democratic design, exercised in the participatory design movement and its later developments into Human Centered Design and HCI.
Through these connections we can re-specify the otherwise overwhelming issue of how AI will impact society and turn this nexus of questions into manageable and timely research topics.
Approach
Our research focusing on Human Centered Machine Intelligence addresses the design of systems and the design of skills, as well as the organizational contexts in which such systems are implemented.
Central to this approach are questions relating to how to design machine learning systems so to operate in the context of human users. How are we to create systems that retain some sensitivity to the cognitive, communicational and organizational practices in which they are embedded? In what ways could the scientific understanding of such processes and practices also help inform the deployment of new systems? These questions tie in with issues such as transparency, explainability, and accountability, all of which are important contemporary topics of AI research.
However, it is probably not enough to set up systems that understand humans and that help the users understand these systems. People also need some basic training in computer science and AI. Without this, the inner workings of intelligent systems, the ways they solve problems and learn from data, risk taking on an alien or magical appearance.
Learn to intelligently interact with complex systems
In order to enable critical scrutiny, part of our approach is directed at how to design learning trajectories of humans so that they can intelligently interact with increasingly complex semi-autonomous systems. This inquiry opens up a range of questions: In what ways should domain expertise interact with technical expertise in order to build new systems? What kinds of interdisciplinary overlaps are needed? How should we understand the requirements for a sustainable system? What domain expertise is required for its maintanence and its further development? Are these crucial competencies being replenished with new generations of users, or are they gradually exhausted from a pre-existing pool?
Mitigate the unintended consequences of this disruptive force
By looking at the effects of Artificial Intelligence on the transformation of society, from the approach of Human Centered Machine Intelligence, the department aims to contribute to the future development and help mitigate some of the unintended consequences of the disruptive force of the technology.