Ground Truths without Grounding: Ethical Considerations for the use of Computer Perception to Help Establish Neural Signatures of Illness
Saturday, September 21, 2024
8:45 AM – 9:45 AM CT
Location: Midway 6 (First Floor)
Abstract: Neurorights scholars are raising ethical concerns around non-clinical uses of neural data, e.g., workplace surveillance or as a form of scientific evidence of brain state or functioning in legal settings (to help infer a defendant's emotional or mental state at the time a crime was committed). Neuroethicists have advocated for the right to neuroprivacy and neurodiversity and cautioned against identifying neural markers for non-clinical applications, e.g., to identify “criminality,” violence, or impulsivity. Here, we draw attention to pioneering methods in neuromodulation to identify neural “signatures” of psychiatric/neurodegenerative diseases which rely on linking intracranial neural data with established clinical measures. These approaches are increasingly being supplemented by data from computer perception (CP) technologies (e.g., passively collected data from wearables; computer vision), which are intended to provide additional insights into mental and behavioral states. While these (primary) avenues of research hold great promise for addressing clinical conditions, they open the door to unintended (secondary) uses of such data, including decision-making for non-clinical applications with high stakes repercussions. Drawing on qualitative insights (n=100 stakeholder interviews) from an NCATS-funded, multisite study of ethical issues around CP, we argue that both primary and secondary uses of neural signatures are critically limited by the current unreliability of CP technologies as objective indicators of subjective states. We discuss why this is the case and propose 1) what work must be done to better achieve the goal of identifying neural markers of brain states and 2) what ethical concerns are raised if and when this goal is achieved.
Learning Objectives:
After participating in this conference, attendees should be able to:
understand stakeholder (clinician, developer, ethics/legal/policy scholar, patient, and caregiver) concerns surrounding the use of computer perception technologies as objective indicators of subjective states
evaluate the limitations of using computer perception to establish neural signatures of brain states
discuss ethical concerns raised by the use of computer perception-informed neural signatures in non-clinical applications
Anika Sonig – Baylor College of Medicine, Center for Medical Ethics and Health Policy; John Herrington – Children's Hospital of Philadelphia; Eric Storch – Baylor College of Medicine; Gabriel Lázaro-Muñoz – Harvard Medical School, Center for Bioethics; Kristin Kostick-Quenet – Baylor College of Medicine, Center for Medical Ethics and Health Policy