> ## Content Index
> Fetch the complete content index at: https://www.fdaweb.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Congressional Report Outlines FDA Oversight of AI-Enabled Devices
- URL: https://www.fdaweb.com/congressional-report-outlines-fda-oversight-of-ai-enabled-devices/
- Published: 2026-06-11T12:00:00.000Z
- Updated: 2026-09-14T13:41:00.000Z
- Author: David McFarland
- Tags: Drugs, #legacy-id-D5161298

A new Congressional Research Service (CRS) report outlines FDA’s evolving approach to regulating artificial intelligence (AI)-enabled medical devices and highlights ongoing debate over whether the agency's existing authorities are sufficient to oversee rapidly advancing AI technologies.

The [6/10 report](https://www.congress.gov/crs%5Fexternal%5Fproducts/IF/PDF/IF13245/IF13245.1.pdf?ref=fdaweb.com) comes as AI applications become increasingly common across healthcare settings, ranging from administrative and financial functions to clinical decision-making tools. While many AI-powered applications fall outside FDA oversight, the agency regulates software functions that meet the statutory definition of a medical device under the Federal Food, Drug, and Cosmetic Act.

According to CRS, FDA has authorized approximately 1,450 AI-enabled medical devices since the first such product was cleared in 1995, with the largest concentrations in radiology, cardiology, and neurology. Most have entered the market through the agency's 510(k) clearance pathway for moderate-risk devices.

The report notes that FDA has not yet authorized a generative AI-enabled medical device, although the agency granted breakthrough device designation in March 2026 to a patient-facing clinical generative AI application developed by RecovryAI. FDA has also convened its Digital Health Advisory Committee to examine regulatory issues surrounding generative AI technologies.

A central challenge identified in the report is that AI-enabled products are expected to evolve after deployment, whereas FDA's traditional device framework is largely designed around products that remain relatively static after clearance or approval. Historically, significant postmarket changes have required manufacturers to submit new regulatory filings, such as updated 510(k) notifications or premarket approval supplements.

To address that issue, FDA proposed a framework in 2019 allowing manufacturers to submit predetermined change control plans (PCCPs) that outline anticipated future modifications and how they will be evaluated. Congress subsequently authorized PCCPs through the Food and Drug Omnibus Reform Act of 2022, allowing certain AI-enabled devices to undergo predefined updates without requiring new premarket submissions, provided safety and effectiveness are maintained. FDA finalized guidance on PCCPs for AI-enabled devices in 8/2025.

Cybersecurity has emerged as another key regulatory concern, according to CRS. Many AI-enabled devices meet the statutory definition of a "cyber device" because they incorporate software and connect to the internet. Manufacturers must demonstrate cybersecurity protections and establish processes to identify and address vulnerabilities throughout a product's lifecycle. FDA issued final cybersecurity guidance in February addressing these requirements.

The CRS report also highlights growing disagreement among policymakers and stakeholders regarding whether FDA requires additional statutory authority to regulate AI products effectively. A 2024 Government Accountability Office report recommended that FDA identify and communicate any legislative changes needed to oversee AI and machine learning-enabled devices.

Congress has separately directed FDA to assess whether new authorities are necessary to monitor the post-deployment performance and patient safety impacts of AI-enabled devices and to report its findings after enactment of the agency's fiscal 2026 appropriations legislation. FDA has already begun exploring real-world performance monitoring approaches through a 2025 public comment process.

The report notes that generative AI products may present particularly difficult regulatory questions because of their potential to generate inaccurate information, perform inconsistently across environments, and rely on opaque training datasets. Some stakeholders have argued that entirely new regulatory frameworks may be required for generative AI systems, with some suggesting such technologies may more closely resemble healthcare providers than traditional medical devices. Others contend FDA's existing authorities provide sufficient flexibility to regulate the technology as it evolves.