WASHINGTON, July 13 — The U.S. Food and Drug Administration is facing growing pressure to modernize how it oversees artificial intelligence in medical devices, as the number of AI-powered tools in American hospitals continues to climb.
As of 2023, the FDA had authorized more than 500 AI and machine-learning-enabled devices. Think of it like a medical toolkit that keeps getting new smart tools — except the tools themselves can change how they work after they’re already in the doctor’s hands.
The vast majority of these authorized devices — over 75% — are in radiology. That means AI is most commonly used today to help radiologists spot tumors, fractures, or other abnormalities in medical images. In plain terms, a computer is looking at your X-ray alongside a human doctor, not instead of one.
The FDA classifies these technologies as “software as a medical device,” or SaMD. In 2021, the agency released an action plan for AI and machine-learning-based SaMD that focuses on three areas: transparency, real-world performance monitoring, and algorithmic bias. The practical upshot is that the FDA wants to know not just whether an AI works in a lab, but whether it works fairly and accurately across different patient populations over time.
Then in 2022, the FDA issued draft guidance on something called predetermined change control plans. Here’s what that means: normally, if a company wants to update a medical device, it has to go back to the FDA for a new review. But AI algorithms can learn and adapt after they’re deployed — they can improve or change on their own. The draft guidance proposes a way for companies to plan for those changes in advance, so the device can evolve without needing a new approval for every tweak.
But there are limits. The FDA has not yet approved any fully autonomous AI diagnostic system — meaning no AI that operates without human oversight. For now, every AI tool in American medicine still requires a human doctor in the loop.
Critics argue the current regulatory framework is insufficient for algorithms that can change rapidly. The Government Accountability Office weighed in during 2022, recommending that the FDA develop a more comprehensive approach to AI oversight. The GAO’s recommendation essentially says the agency’s existing tools weren’t built for software that learns on the job.
So what’s next? The FDA’s 2021 action plan and the 2022 draft guidance on change control plans suggest the agency is moving toward a more flexible system. But the gap between the 500-plus authorized devices and the zero fully autonomous systems shows just how carefully regulators are treading. For patients, the immediate takeaway is that AI is already part of modern medicine — but always with a human doctor making the final call.
Consult your doctor for medical advice.
Sources
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “The FDA and AI Regulation: Ensuring Safety in Medical Algorithms”, “description”: “WASHINGTON, July 13 — The U.S. Food and Drug Administration is facing growing pressure to modernize how it oversees artificial intelligence in medical devices, as…”, “datePublished”: “2026-07-13T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Erika Johnson”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.2bme.xyz/”}, “citation”: [{“@type”: “CreativeWork”, “url”: “https://en.wikipedia.org/wiki/Food_and_Drug_Administration”}]}
