Written by Yiya Wang, Class of 2028
Congratulations to Prof. Daniel Weissglass for publishing a paper “AI preference prediction beyond substituted judgement: enhancing best interest decision-making” in the leading medical ethics journal, Journal of Medical Ethics, co-authored with current and former DKU students, with the entire project originating from a regular classroom course assignment.
It is the product of a research-oriented teaching method Prof. Weissglass uses in applied ethics courses, which asks students to work as part of a research team with Prof. Weissglass in exploring cutting-edge ethical questions. This paper began in global health ethics, and was supported in further development by the SRS program. It resulted in a presentation at a professional workshop in Singapore before ultimately being prepared for successful publication in the Journal of Medical Ethics, which is arguably the most impactful journal in the field. It includes six coauthors who were enrolled in that course – some of whom are still at DKU.
As a representative of the student co-authors, Xinyu Zhou shared her experience and insights, highlighting the contributions and the valuable gains from this collaborative project.
The project began as a group assignment in Prof. Weissglass’s Global Health Ethics course, with no initial expectation of publication. She and her team chose ICU decision-making as their topic, drawn to the ethical complexities of incapacitated patients unable to express their preferences. They spent weeks refining the vague initial topic to lay the groundwork for their research.
Xinyu Zhou faced key challenges evaluating AI patient preference prediction (AIPP) models ethically, struggling with tensions around accuracy, population-level data use and patient autonomy. With weekly guidance from Prof. Weissglass, the team shifted their focus, from defending AIPP accuracy to exploring its value despite ethical criticisms, breaking the research deadlock.
This shift reshaped her understanding of research that high-quality research is iterative, not linear, and rethinking logic amid obstacles drives progress.
As a co-author, her experience exceeded typical RA work that she contributed ideas from the project’s start, collaborated closely with Prof. Weissglass, and actively participated in drafting and revising the manuscript.
Through this, she distinguished formal research from coursework, learning it requires engaging with literature, anticipating counterarguments, and responding to peer feedback, making her feel part of global academic dialogue.
The project also transformed her perception of research as a collaborative, iterative process, building her academic literacy, and fostering flexible critical thinking to navigate bottlenecks.
Additionally, the experience shaped her career plans, deepening her passion for healthcare ethics and solidifying her goal to work at the intersection of healthcare and analytics, with practical insights to prioritize patient-centered outcomes.
Abstract
Tracking patient preferences is vital to medical decision-making, but evidence suggests that the standard method for tracking the preferences of incapacitated or incompetent patients (ie, surrogates) is inaccurate. Recent proposals suggest that artificial intelligence preference predictors (AIPPs) can improve preference tracking for these patients, but have faced significant objections. While many of these objections depend on unsettled empirical or technical assumptions, one prominent objection—that AIPPs rely inappropriately on impersonal information—seems to be an in-principle challenge to AIPPs. In this paper, we show that even granting an implausibly strong version of this objection, AIPPs may provide value to clinical decision-making. To show this, we develop suggestions that AIPPs may support best interest decision-making (BIDM) by improving the accuracy, consistency and speed of BIDM, and show that the prevalence of BIDM in the intensive care unit (ICU) grants this application of AIPPs significant moral and practical consequence. This not only clears a path to improve BIDM but also establishes a safe harbour—a relatively uncontroversial yet impactful space—in which proponents may develop AIPPs sufficently to resolve empirical and technical questions about their potential. We conclude by highlighting key questions for the application of AIPPs to BI determinations, setting an agenda for the deeper examination of a largely overlooked application of these tools.