Enabling Domain Expert Evaluation of Emerging AI Technologies in Healthcare Settings
Contact: Shatha Degachi (C.Degachi@tudelft.nl)
AI technologies have been evolving at a rapid rate in the past few years, and have similarly seen a growth in adoption across sectors of industry and public life. One such sector is healthcare. The healthcare setting brings many challenges to the prospect of AI adoption with its high requirements for trust, safety, and reliability, thus it is vital to enable healthcare experts to understand and make decisions about the feasibility of adapting AI tooling to their needs. In this project, we use the concept of imaginaries to investigate health expert mental models of AI, and how they would assess future technologies with AI-like properties (non-determinism, low transparency, data fragility), and see beyond the hype. We would develop a set of recommendations from this work which enable AI developers, medical experts, and regulators to communicate expectations and requirements better and make decisions accordingly.
