The world of artificial intelligence (AI) regulation is still in its infancy, but a number of agencies are nonetheless keen on harmonization for at least some of this policymaking task. The FDA announced recently that it has joined with Health Canada and the U.K. Medicines and Healthcare Products Regulatory Agency to develop a series of 10 guiding principles for good machine learning practices (GMLP), thus answering one of the key questions facing developers of these algorithms.

There are few guidelines of any sort that are specific to artificial intelligence (AI) for medical devices, but that doesn’t mean there are no signposts for developers. There are existing product marketing authorizations that offer some insights, but the FDA’s Bakul Patel said a risk stratification guidance by the International Medical Device Regulators Forum (IMDF) is an example of a non-AI blueprint for how the FDA will ultimately approach regulation of AI.

The U.S. FDA has issued an action plan for regulation of artificial intelligence and machine learning (AI, ML), which includes issuance of a draft guidance for change control for adaptive algorithms. There is no guarantee a final guidance will emerge before 2022, however, leaving developers with another year – perhaps longer – of uncertainty as to how to handle change control for their algorithms.