Applications and Challenges of Human Computer Interaction and AI Interfaces for Health Care

Artificial intelligence (AI) based systems have accelerated in popularity, performance, and complexity over the past several decades. This is especially true in the health care setting, where intelligent systems are capable of informing medical decision making. To ensure that emerging AI systems are clinically useful, ethical, and safe, it is crucial to leverage principles from human computer interaction (HCI). By reviewing the historical and present-day applications of systems that benefited from the synergistic coupling of HCI and AI, we aim to emphasize the critical need for continued HCI and AI unification (HCI-AI). Specifically, we will highlight the evolution of HCI and AI, which both emerged during the mid-twentieth century. During this time, the fields of HCI and AI largely operated separately from one another due to contrasting principles and conflicting ideas of where the human stood in relation to machines. Over time, this early division healed, and intelligent medical systems increasingly evolved to incorporate HCI principles to help physicians more accurately diagnose and treat patients. With the expansive growth of HCI and AI technology in the twenty-first century, there have been an increased number of efforts to implement HCI-AI systems that can support a myriad of health care problems spanning the clinical, public health, and biomedical research settings. Despite the expansion of HCI-AI for health care applications, barriers to successful implementation continue to persist. We discuss current efforts, policies and guidelines intended to mitigate these challenges. We conclude by recommending proactive development of HCI-AI systems which leverage the strengths of both human and AI to develop systems that can continuously improve and remain clinically useful and safe.