The RCMI Program in Health Disparities Research at Meharry Medical College

Meharry Medical College, Nashville TN

Investigators

Linked publications & trials

Abstract

ABSTRACT Identifying and addressing ethical AI challenges is critical in healthcare domain as AI has achieved human-level performance in skin cancer classification, diabetic retinopathy detection, chest radiograph diagnosis, sepsis treatment nowadays, however not too many clinical AI solutions are deployed in hospitals or actively utilized by physicians. The major problem is that existing clinical AI methods are less trustworthy, non- explainable, suffer with algorithmic and societal biases, and have security and privacy issues which leads to misinformation, misdiagnosis, maltreatment, and disparities. Thus, there is a great need to promote awareness about ethical AI to healthcare professionals (AI algorithm user) and data scientists/computer scientists (AI algorithm developer) to responsibly develop and deploy AI in healthcare. We hereby submit an Administrative Supplement in response to the Notice of Special Interest (NOSI) NOT-MD-23-013. In order to better position Meharry for the challenges posed by the algorithmic biases, transparency, explainability, and inherently transdisciplinary nature of health disparities, this RCMI administrative supplement will focus on identifying and addressing ethical considerations in AI/ML technologies in healthcare. The supplement project will also develop strategies for training the investigators, staff, medical professionals, and graduate students regarding potential risks, opportunities and challenges for ethical and responsible AI in healthcare, at the NIMHD-funded RCMI Program in Health Disparities Research at Meharry Medical College. Therefore, in response to the NOSI (NOT-MD-23-013), this request for the Administrative Supplement will specifically support collaborations between SACS, RCMI investigators, and medical professionals to identify and address the ethical considerations of applying AI/ML technologies in healthcare for building trustworthy AI/ML applications. To promote awareness around responsible usages of AI/ML in healthcare this Administrative Supplement will specifically focus on two specific aims indicated below. Aim 1: To identify and address ethical considerations in AI/ML Technologies in healthcare: This aim will focus on identifying and mitigating bias in the data and AI/ML models developed in the healthcare domain. Aim 2: Enhance Meharry’s RCMI capacity by providing ethical and responsible AI training to RCMI investigators, medical professionals, staff, and graduate students: This aim will focus on enhancing Meharry’s RCMI capacity by providing ethical and responsible AI training to RCMI investigators, medical professional, staff, and graduate students through a series of virtual tutorial sessions, SACS professional development training in AI ethics and incorporating the ethical AI concepts in the existing courses. In summary, the NOSI is aligned with the primary goals of the NIMHD funded RCMI center’s vision “to focus on the underserved communities health and reduce health disparity through technology and innovation”. The proposed project will help in identifying and addressing the algorithmic and societal bias in the healthcare datasets generated and AI/ML models built using those datasets. In addition, the training on trustworthy and responsible AI will promote awareness among health care professionals, and RCMI investigators for using ethically sound practices, and develop strategies to reduce health disparity.