12 Plagues of AI in Healthcare: A Practical Guide to Current Issues With Using Machine Learning in a Medical Context

Abstract

The healthcare field has long been promised a number of exciting and powerful applications of Artificial Intelligence (AI) to improve the quality and delivery of health care services. AI techniques, such as machine learning (ML), have proven the ability to model enormous amounts of complex data and biological phenomena in ways only imaginable with human abilities alone. As such, medical professionals, data scientists, and Big Tech companies alike have all invested substantial time, effort, and funding into these technologies with hopes that AI systems will provide rigorous and systematic interpretations of large amounts of data that can be leveraged to augment clinical judgments in real time. However, despite not being newly introduced, AI-based medical devices have more than often been limited in their true clinical impact that was originally promised or that which is likely capable, such as during the current COVID-19 pandemic. There are several common pitfalls for these technologies that if not prospectively managed or adjusted in real-time, will continue to hinder their performance in high stakes environments outside of the lab in which they were created. To address these concerns, we outline and discuss many of the problems that future developers will likely face that contribute to these failures. Specifically, we examine the field under four lenses: approach, data, method and operation. If we continue to prospectively address and manage these concerns with reliable solutions and appropriate system processes in place, then we as a field may further optimize the clinical applicability and adoption of medical based AI technology moving forward.

Prediction of orthognathic surgery plan from 3D cephalometric analysis via deep learning

“…For an AI-based prediction model in medicine, clinicians expect more than to transfer valuable decision-making process of experienced surgeons to the model. The ability to present larger amounts of interpretable information is of great importance to augment surgeons' clinical judgements and gain their trust [ 13 ]. When the diagnosis or treatment planning are inconsistent among different doctors, it is important that an AI-based model can provide valuable information to assist decision and decrease biases [ 14 ].…”

Parcellation‐based tractographic modeling of the salience network through meta‐analysis

“…However, the current work provides a more detailed cortical model of the ROIs and connections involved in the SN by utilizing combined structural and functional neuroimaging studies in the literature and by describing our results according to the detailed HCP parcellation scheme (Glasser et al., 2016 ). Unfortunately, the structural interconnectedness of the SN has previously remained underspecified despite both its increasing body of research over previous years and the large advancements in neuroimaging technologies made in the neuroscience community (Doyen & Dadario, 2022 ; Menon, 2015 ). Such precision and clarity of the structural white matter connectivity of the SN are necessary to better understand the essential functions of the SN according to individual neural substrates and how to navigate this region with clinical applications (Menon, 2011 ; Rosen et al., 2021 ).…”

Model Blindness II: Investigating a Model-Based Recommender System’s Impact on Decision Making

“…Artificial Intelligence (AI) techniques, like machine learning, have increased in popularity because of their proven ability to model large amounts of complex data and biological phenomena. Consequently, big tech companies, data scientists, medical professionals, and nearly every large workplace is investing more resources to deploy these technologies (Doyen & Dadario, 2022). However, deploying these technologies without proper evaluation in complex socio-technical environments can have enormous consequences for end users within these organizations.…”