FDA Sets Eyes on Regulating AI in Healthcare

AI

The US Food and Drug Administration wants to make regulations in using Artificial Intelligence in the healthcare industry / Photo by: The U.S. Food and Drug Administration via Flickr

 

The Food and Drug Administration has issued Tuesday a framework that aims to guide the use of artificial intelligence (AI) in the medical industry. It is part of the agency’s initiative to provide better insight into this budding technological use.

As reported by Stat News, the FDA will be making the framework in an attempt to regulate AI products that rely on obtained new data and are currently being used in medicine. According to FDA’s outgoing commissioner Scott Gottlieb, the “rapidly evolving segment of AI products” would eventually need some looking after, as the agency reveals that when AI takes over in the healthcare industry, the FDA would have more things to deal with on their plate.

So the answer is to regulate them early on to ensure that proper oversight is exercised.

The white paper containing the criteria would look at medical products that rely on AI. From these, the FDA will examine which products would need a review before releasing it. The review will look at the performance of the product’s algorithm, “a manufacturer’s plan to make modifications, and the manufacturer’s ability to manage the risks associated with any modifications.”

Said Eric Topol, an expert in artificial intelligence at the Scripps Research Institute, the white paper is a noteworthy forethought that is just what this new innovation needs, especially since the world of connectivity we have today only creates a climate that encourages the use of more and more technology.

As of late, the early framework will still need crucial input from the public and some stakeholders. After that is completed, the FDA then intends to push through with understanding policies that would be better suited for the regulation of adaptive AI systems.

The FDA has already begun work on regulating medical devices with locked algorithms, devices that are not under the mercy of the ever-changing algorithm.