AI and Machine Learning in IBD: Promising Applications and Remaining Challenges

Nearly 1 in 100 Americans have Inflammatory Bowel Disease (IBD), with up to 56,000 new cases being diagnosed each year.¹ IBD is a complex disease with a myriad of presentations, possible treatment approaches, and patient outcomes. Artificial intelligence (AI)—a field of technology which began in the 1950s—refers to the ability of computers to learn and perform tasks that would have typically required human intelligence, while “machine learning” refers to the development of the algorithms that help AI learn patterns from data.^2,3 The goal in many industries, including health care, is for AI to aid in and improve decision-making. Applications of AI including machine learning already greatly influence the oncology space, aiding in risk assessment, early diagnosis, prognosis, and treatment decision-making.⁴ Similar utilizations are being investigated to help improve the quality and efficiency of care for patients with IBD, but there is still much research to be done before we can fully leverage such tools in everyday practice.⁵

Although extensive progress in AI has been made since the turn of the century, several limitations remain. Poor-quality data sets may lead to inaccurate predictions, and it is difficult to generalize data sets to minority populations. In health care, clinicians must also understand and be able to interpret the algorithms in order to trust and apply them in practice. Lastly, and importantly, there are ethical concerns regarding patient privacy in data collection.⁶

1