The network formed by the Trustworthy AI labs allows for a wide range of specialized experts to contribute to the development of AI projects conducted by academia, government bodies, municipalities, or industries, assessing whether these projects are ethically sound and reliable.
The assessment process utilizes an applied ethics-based approach called Z-Inspection which is based on the EU's definition of what constitutes trustworthy AI.
- Z-Inspection is a general inspection process applicable across various domains where AI systems are utilized, such as business operations, healthcare, and the public sector, among others. Z-Inspection was the pioneering process in assessing trustworthy AI in practice and, as far as we know, remains the most comprehensive method available, explains Dr. Pedro Kringen.
He is one of the initiators behind the Trustworthy AI lab and the developer of the Z-Inspection method. Professor Roberto Zicari, who is the initiator of the entire global project, is also a member of the board of the Trustworthy AI Lab Norway. Currently, there are 25 Trustworthy AI labs established worldwide.
"Together, we might contribute to better and safer applications of artificial intelligence, where ethical rules and norms are followed, while these applications contribute to a better society."
Frode Ramstad Johansen, Associate Professor at Østfold University College
Better and safer AI
The international AI network, which the Norwegian university college is now joining, was initiated by researchers from various fields at the Big Data Lab at Goethe University in Frankfurt.
At Østfold University College, the start-up for the trustworthy AI lab was initiated by innovation researchers. They now aim to involve colleagues from diverse disciplines, as well as external partners, to create the best possible foundation for actively and broadly advocating for a conscious use of artificial intelligence.
- The method is interdisciplinary, so a successful lab requires diverse expertise. So far we have gathered expertise from innovation, organizational development, law, ethics, medicine, and health, but we aim for a broader expertise, says Associate Professor Frode Ramstad Johansen from the Department of Engineering.
- What are your hopes for this initiative, and how important is it from an AI perspective?
- We aim to create a forum for applied AI projects at the university, connected to the international research community through the other 25 AI labs globally. Together, we hope to contribute to better and safer applications of artificial intelligence, adhering to ethical rules and norms, while also contributing to a better society," says the HiØ researcher.
Starting the new year with a pilot project
According to Ramstad Johansen, the Trustworthy AI labs are not labs in the traditional, physical sense, but rather interdisciplinary and international forums that encourage debate and reflection on the responsible use of artificial intelligence. These labs offer events and opportunities for teaching and research.
During the opening meeting recently held at the campus in Fredrikstad, Ramstad Johansen and Pedro Kringen shared information about the possibilities of the lab concept with both HiØ staff and external guests. Now that the lab has been established the work begins in earnest.
- We have two main tracks ahead. One is to build up the lab with interested resource persons and spread knowledge about the method. The other priority is preparations for an international pilot project assessing the use of generative AI in higher education, Ramstad Johansen says.
The pilot project at HiØ is set to begin in early 2024.
Tailored for the Norwegian AI market
According to Kringen so far only academic institutions have been involved in establishing Trustworthy AI Labs.
- This is because we have maintained this as a purely academic exercise, and all work is carried out pro bono, he says, adding:
- However, several labs have applied for and received extensive support for the use and development of the method.
Dr. Kringen emphasizes that artificial intelligence is rapidly evolving, raising concerns about the ethical and societal implications. He says all researchers involved in the international lab community are passionate about the issue and aim to make a difference.
The purpose of establishing a dedicated lab in Norway is, among other things, to actively participate in developing methods for assessing artificial intelligence that meet specific Norwegian needs.
- The research community at HiØ is diverse and has strong connections to the business sector and the municipality, providing unique opportunities, says Kringen.
Ramstad Johansen hopes that the new AI lab can serve as a common platform for AI-related projects and initiatives at Østfold University College.