Essential Concepts in Artificial Intelligence: A Guide for Pediatric Providers

TL;DR

This article explains key AI concepts for pediatric providers, showing how AI is changing healthcare and offering tools to assess and develop AI models.

Contribution

The article introduces an AI model checklist and discusses challenges in pediatric AI development with potential solutions.

Findings

AI is increasingly used in cardiology and radiology with approved medical applications.

Pediatric AI development faces unique challenges compared to adult-focused tools.

Multisite collaboration methods can help overcome barriers in pediatric AI development.

Abstract

Artificial intelligence (AI) has exploded in public awareness over recent years and is already beginning to reshape the health care sector. Yet, even as AI becomes more prevalent, it remains a mystery to many providers who lack hands-on exposure during their training or on the job. Intended for medical professionals, this article defines essential concepts in AI interspersed with illustrations of how such concepts may be applied within cardiology and radiology—fields that have garnered the most approved medical AI applications to date. No experience in the field of AI is requisite before reading. To assist providers encountering novel machine learning tools, we also present an AI model checklist to empower critical assessment. We finally discuss hurdles in the path of developing pediatric AI tools—including challenges distinct from the adult setting—and discuss potential solutions, including various methods of multisite collaboration. This article aims to increase the engagement of health care professionals who may encounter AI models in practice or who seek to become involved in AI development themselves. We encourage the reader the freedom to either peruse this article in its entirety or to reference specific concepts individually. Terminology central to machine learning is emphasized in bold.

Figures