Ethical Issues Using AI in the Field of Pediatrics

Abstract

The gap between AI applications in adult and pediatric medicine remains large as children illustrate unique developmental and physiological factors compared to adults. These distinctions restrict the amounts of data available, preventing swift and sufficient training and functionality for many AI algorithms. Several adult applications in radiology and gastroenterology have been analyzed for potential use in pediatrics; however, many cannot be directly translated or applied. Factors regarding ethics and related concerns also limit applications from being rapidly employed. This chapter explores the ethical and practical issues that surround the management and implementation of AI in pediatric medicine. An array of concerns arise with big data, including complications, potential risks, and future safety of children. Additionally, the general uneasiness of how, what, and where data is utilized naturally prevails with parents as primary decision-makers. Parental concerns mainly focus on these data management issues as well as cost, convenience, absence of human care, and accuracy of AI algorithms. The latter concerns are often considered when AI algorithms must be carefully implemented for special diagnoses. For example, Telehealth and chatbots are cost-efficient and convenient, but may not be appropriate modalities for optimal treatment plans and clinician–patient relationships. Online modalities can mimic human connection but many fear that it only prevails to a limited extent. This chapter also offers insight on current and emerging AI applications in diverse subspecialties in pediatrics such as radiology, cardiology, and gastroenterology. The methods and networks utilized in those AI applications will be analyzed to review their potential in pediatric medicine. Most current applications exist more as an AI-assist tool rather than an AI-replace tool. Numerous tools assist the clinician by enhancing accuracy and aiding with identifying diseases. These conditions encompass common diseases—diabetes mellitus and pneumonia—and rarer illnesses like cancer and irritable bowel disorder (IBD). With AI being gradually incorporated in medicine, trust for both clinicians and patients are another factor to consider. Trust is affected by elements like education and prior biases. Training for health professionals for AI applications will also be reviewed as training can affect optimal treatment options. Review of ethical elements, concerns, and factors for AI implementation in this chapter will help illustrate the potential of AI in pediatrics.

Ähnliche Arbeiten