Research Issues and Challenges in the Computational Development of Trustworthy AI

The development and deployment of AI systemsnecessitate a steadfast commitment to reliability, safety,security, ethics, and social responsibility. This paper introduceskey research issues and challenges for trustworthy AI based onour experience working on several ongoing research projects atMälardalen University (MDU), Sweden, which considerspractical, real-world scenarios from the mobility,transportation, and healthcare domains. Our observationshave highlighted several critical technical components thatunderpin trustworthy AI. These components include fairness,safety, transparency, explainability, accountability, rigoroustesting, verification, and a human-centric approach to AI.Notably, these elements align closely with the current state-ofthe-art practices in the field.