Food and Drug Administration prompted take care of the concerns about the algorithms’ lack of transparency and potential bias affecting patient outcomes. Food and Drug Administration has taken a multipronged approach to the problem as artificial intelligence and machine learning are increasingly used in medical devices. In January, the Food and Drug Administration released an initial strategy that outlined five steps the agency intends to pursue.
A transparent, patient-centered approach, creating new pilot projects to enable real-world performance monitoring, and constructing a regulatory framework are among them. The capacity of AI/ML in software to learn from real-world use and experience and improve its performance is one of the most significant advantages. However, as a Harvard T.H. Chan School of Public Health report published in October 2020 found, there are still concerns about data transparency, including how it is obtained, the general quality of the data, and how it is validated.
The potential value of AI/ML-based SaMD rests in its ability to continually learn, as the agency pointed out in an April 2019 discussion paper. In the report, the adaption or change to the algorithm is realized after the SaMD had been released for usage. It has also learned from real-world experience.