Search

This article identifies and examines a tension in mental health researchers' growing enthusiasm for the use of computational tools powered by advances in artificial intelligence and machine learning (AI/ML). Although there is increasing recognition of the value of participatory methods in science generally and in mental health research specifically, many AI/ML approaches, fueled by an ever-growing number of sensors collecting multimodal data, risk further distancing participants from research processes and rendering them as mere vectors or collections of data points. The imperatives of the "participatory turn" inmental health research may be at odds with the (often unquestioned) assumptions and data collection methods of AI/ML approaches. This article aims to show why this is a problem and how it might be addressed. [ABSTRACT FROM AUTHOR]

The authors comment on an article by I. G. Cohen on what ChatGPT means for bioethics. They discuss one shortcoming of Cohen's approach, the treatment of ChatGPT as the paradigmatic case of cutting-edge AI which may stretch the limits of bioethics, and examples of what are expected to be a growing research program that harnesses the power of transformers in conjunction with the rich multimodal datasets generated in the medical domain.