Artificial intelligence used in healthcare requires a distributed regulation approach that involves both centralized agencies like the FDA and decentralized agencies, argues a team of researchers in a May 26 PLOS Digital Health article.
The group of researchers from Harvard University, Idaho State University and U.K.-based Imperial College argues that given some of the challenges health AI presents, like the black box issue of determining causality, the technology is in need of regulation. The authors reason that this regulation should overwhelmingly come from local health systems as opposed to federal bodies.
Centralized agencies like the FDA do not have the necessary resources or capacity to analyze health AI and the scale that would be required, the authors argue. They also say that given that medical professionals will be the ones in charge of using such technology, they should have a hand in the regulation and decision-making process, something that would be difficult to achieve through centralized regulation alone.
"It is now up to existing regulatory organizations to propose an operational plan to realize a fairer and more effective distributed approach to the regulation of clinical AI," they write.