ELSIcon2024: Artificial intelligence for precision medicine: How close to ‘personalized’ can we really get?

ELSIcon2024 • Panel • June 11, 2024 

Authors:

Corresponding Author: Carole Federico, MSc, PhD (she/her/hers) – Stanford University

Panelist: Kristin Kostick-Quenet – Baylor College of Medicine

Panelist: Nicole Martinez-Martin, JD, PhD (she/her/hers) – Stanford University

Panelist: Abdoul Jalil Djiberou Mahamadou, PhD (he/him/his) – Stanford University

The use of artificial intelligence (AI) to facilitate prediction and decision-making has become widely popular across many disciplines. Fueled by advances in computing power, algorithms, and big data, the last decade has seen the widespread application of AI to diverse fields, including health care. The convergence of AI and precision medicine promises to revolutionize health care by enhancing disease prevention, and personalized diagnosis and treatment. For example, AI can be used to identify combinations of therapeutic targets and mechanisms in order to develop treatment strategies for complex diseases. Beyond clinical care, AI-enabled precision medicine offers the possibility of vast personalization of AI tool interfaces and user experiences. There is little doubt that AI presents tremendous opportunities to achieve the goals of precision health care. Questions remain, however, about what it will really take to personalize medicine, and whether computers can accurately and effectively account for the vast amount of heterogeneity present in patient data, social determinants of health, patient illness trajectories, and the diverse meaning and significance of symptom experiences. While sophisticated AI methodologies operating on large, diverse datasets can move us closer to this goal, new types of longitudinal, multimodal data will be required before medicine can be truly ‘personalized.’ Satisfying the need for more – and more personalized – data may be accomplished thanks in part to advances in genomic sequencing and digital phenotyping technologies, but doing so raises a number of ethical concerns, including questions about data availability and quality, safety and privacy, and bias and fairness.

At the bottom right corner of the video, you can click the “CC” button to display closed captions. You can also click the icon to the right of the “CC” button to display the transcript.