Regulatory Approach Allowing AI-based Medical Devices Continually Learn

FREMONT, CA: With the continuous evolution of medical technologies, it is now possible to use AI-based medical devices to detect illnesses and symptoms. Cognitive computing cuts time and cost, and also eliminates redundancy, in order to assists physicians and radiologists, accelerating disease diagnosis.

As part of its market authorization procedures, Food and Drug Administration (FDA) is currently conducting reviews on medical devices. AI-based devices have already been permitted by this board to detect diabetic retinopathy and alert suppliers to prospective patient strokes. The accepted AI algorithms are what the FDA calls locked algorithms, meaning they are static on an ongoing basis. Now, the device manufacturer is updating these locked algorithms at regular intervals.

The new legislative structure would enable what the FDA calls adaptive algorithms, or AI algorithms that learn and adapt through real-world use and do not require manual updates. This would be achieved while ensuring medical device security. It is important for the researchers in the field of medicine and engineers to gain insights into how best to regulate software that is not static but continues to learn based on the latest information that it meets. This regulatory ArtificiaI Intelligence frame work will drive healthcare acceptance of AI.

The idea to regulate the AI in medical devices is a beneficial move, changing the face of the industry when it comes to controlling software, and it poses new questions that need to be addressed. Though AI has been around for years in healthcare, in recent past its use has risen exponentially.

FDA is taking a significant first step in developing guidelines for learning software. The nearer the AI instruments are to direct patient care, the more essential it is to guarantee that the products are secure and operate correctly.