Establishing an artificial intelligence (AI) infrastructure is contingent upon processes being conducted within structured regulatory frameworks that adhere to existing best practices. Institutional policies should include the development of an AI governance body with the goals of proactively monitoring deployed AI solutions for deviations in performance, responsibly scaling them, mitigating potential biases, and ensuring the fundamental principles of transparency and accountability are followed. Specific regulatory considerations include ensuring compliance with Food and Drug Administration regulations as well as ensuring data privacy and Health Insurance Portability and Accountability Act compliance. Implementation challenges that must be considered include workflow integration of AI solutions, user acceptance and trust, initiating a change in hospital culture toward embracing new technology, and building a sustainable technical foundation for maintenance and continuous learning. Emphasizing these considerations is essential to fostering compliant and responsible use of AI that maximizes value and quality of care delivered across a health care system. Likewise, a top-down focus on ethics surrounding responsible use of AI, measures toward the reduction in bias, and emphasizing explainability of AI decision-making will be critical for sustainable and trustworthy integration of systems-wide AI solutions. LEVEL OF EVIDENCE: Level V, expert opinion.
Kunze et al. (Fri,) studied this question.