FDA’s new guidelines for AI

AI in medicine enters a new regulatory hase

In March 2026, the Food and Drug Administration released an updated list of 1 451 certified AI-enabled medical devices.

Radiology-dominance that speaks volumes

The most striking insight from the FDA data is that 76% of all approved AI solutions are in radiology. For years, we have observed that radiology offers ideal conditions for AI deployment:

• it operates on highly standardized imaging data,
• relies on repeatable decision-making processes,
• and allows for relatively straightforward validation of results.

However, AI in medicine is no longer confined to radiology. As the field expands, so too does the regulatory approach. In March, a significant shift became visible in how the FDA approaches classification and evaluation. The updated perspective reflects a stronger emphasis on intended use and risk level.

The FDA is moving toward a more flexible framework for non-invasive devices that monitor vital parameters, such as heart rate, blood pressure, physical activity and other basic health indicators. The key condition remains that these devices must not be used for direct medical diagnosis.

As a result, a new category is emerging “medical-grade wellness devices”, solutions that sit at the intersection of healthcare and consumer technology.

In practice, this means that manufacturers can benefit from simplified market entry pathways, without undergoing the full regulatory approval process required for high-risk medical devices provided they clearly communicate the non-medical nature of their products. At the same time, the FDA emphasizes the importance of transparency toward users, including clear communication about algorithmic limitations and the risk of misinterpretation of data.

Consequently, part of the regulatory burden is shifting from pre-market certification to product design, communication and responsible data use in real-world settings.

This represents a bold move and a clear signal of a broader transformation in market logic. The Food and Drug Administration effectively acknowledges that not every technology processing health-related data needs to be regulated as a traditional medical device. As a result, a parallel innovation pathway is emerging one where solutions evolve faster, closer to the user and increasingly outside the traditional healthcare system.

The key question is whether this shift will ultimately prove to be the right one?

[embed]AI-Enabled Medical Devices The AI-Enabled Medical Device List is a resource intended to identify AI-enabled medical devices that are authorized…www.fda.gov

[embed]US FDA to limit regulation of health and fitness wearables, commissioner says The U.S. Food and Drug Administration said on Tuesday that it will limit regulation of wearable devices and software…www.reuters.com

[embed]General Wellness: Policy for Low Risk Devices - Guidance This guidance provides clarity on FDA's compliance policy for low risk products that promote a healthy lifestyle…www.fda.gov

Sivakumar R, Lue B, Kundu S. FDA Approval of Artificial Intelligence and Machine Learning Devices in Radiology: A Systematic Review. JAMA Netw Open. 2025;8(11):e2542338. doi:10.1001/jamanetworkopen.2025.42338