In this NSF-funded project, systems engineers, psychologists, philosophers, clinicians, and computer scientists at Missouri S&T, Saint Louis University, United Network for Organ Sharing (UNOS), and Mid-America Transplant are collaborating to develop an artificial intelligence (AI) decision support tool for transplant surgeons and organ procurement coordinators to increase kidney utilization in the placement process.
To this end, we have 5 primary research questions:
- How can we measure human preferences to customize an AI decision support system’s operation and output?
- How can we aggregate and improve fairness in the AI output while balancing disparate stakeholder viewpoints?
- How does the communication of uncertainty and explainability of an AI model influence performance, user confidence, perceived ease of use, and trust?
- How can AI decision support systems learn and adapt over time to changing contexts and evolving preferences while being transparent and explainable?
- To what extent does the proposed intervention of an AI decision support system increase kidney utilization (by increasing placement of less desirable kidneys)?