How the FDA Is Using AI in Regulatory Review

MedDevice by Design with Mark Drlik and Ariana Wilson
Resources

How the FDA Uses AI in Device Review

YouTube video thumbnail

The FDA AI in device review process is evolving fast—and it may transform how regulatory teams navigate medical device submissions. In this episode of MedDevice by Design, we explore how the FDA is using artificial intelligence internally to accelerate 510(k) reviews, improve predicate searches, and streamline labeling and documentation checks.

FDA’s Internal AI Tools Are Expanding

The FDA has already appointed a Chief AI Officer and is developing internal tools, including a large language model rumored to be in collaboration with OpenAI. Starting in June 2025, these tools are expected to be rolled out across departments. These technologies are designed to support internal staff—not replace them.

Faster Reviews for Common Device Submissions

Class II devices such as blood pressure monitors and surgical gloves may benefit the most from FDA AI in device review workflows. By using trained AI models, reviewers can automate routine predicate searches, validate labeling against regulatory standards, and identify missing traceability in submission documents.

Why This Matters for Regulatory Professionals

While novel and high-risk devices will still require close human evaluation, AI promises to reduce the review burden for standard submissions. This allows regulatory and engineering teams to focus more on innovative product development. Instead of spending weeks preparing repetitive documentation, they can rely on automated feedback to guide improvements faster.

What Comes Next?

There’s still a lot we don’t know, but the FDA is clearly signaling that AI will be part of the future of regulatory review. The ability to streamline standard 510(k) submissions could help improve speed, consistency, and predictability in approvals—benefiting both regulators and medtech innovators.

Nick from StarFish Medical demonstrating antigen detection using a toy antibody model to explain how monoclonal antibodies bind antigens in ELISA diagnostics

Nick and Nigel break down the ELISA assay explained in simple, practical terms using everyday models.

Magnifying glass revealing cracks with the text “This gets missed,” illustrating hidden risks in medical device validation and real-world use.

Ariana Wilson sits down with Mark Drlik to unpack why reprocessing is often one of the hardest challenges engineers face during development.

Thumbnail showing the text “ETO or Radiation?” with a cloud icon representing ethylene oxide sterilization and a radiating burst icon representing radiation sterilization for medical devices.

Nick and Nigel walk through how teams decide between ethylene oxide, E-beam, and gamma radiation sterilization.

Medical device pilot manufacturing workspace showing engineers supporting NPI and scalable production

In MedTech, success rarely comes from invention alone. Plenty of promising technologies make it through verification and early clinical work, only to stall when the team tries to turn them into something buildable.