Key points are not available for this paper at this time.
The growing excitement about the potential positive impact of generative AI (GenAI) solutions in healthcare has been tempered by uncertainty on how to ensure that such solutions are deployed safely and effectively. A core issue is that the current product-centric regulatory oversight model does not apply well to a technology that adapts to the operating environment and becomes enmeshed with medical practice in a non-deterministic manner. This difficulty has led to the development of alternate approaches to foster the responsible deployment of GenAI solutions that are focused on the organizations developing and using the technology, as they are uniquely positioned to prevent and quickly address issues as they occur. One emerging approach is via the certification with the ISO 42001 standard, which defines the structure for auditable AI Management Systems (AIMS) in organizations developing and/or deploying AI solutions. The process of ISO 42001 certification can enable healthcare organizations to build adaptive and auditable AIMS that mitigate risks while supporting the deployment of trustworthy AI. The wide adoption of ISO 42001 certification by healthcare organizations would allow for the utilization of beneficial GenAI solutions while potentially facilitating the performance of oversight functions by regulators and payers. This article introduces the ISO 42001 implementation process in healthcare organizations and describes the next steps in the operationalization of this new GenAI risk mitigation approach.
Thiers et al. (2024) studied this question.