Advisories April 22, 2024

FDA/Food, Drug & Device Advisory: How the FDA Is Keeping Up with Artificial Intelligence and Machine Learning in Drug and Device Product Development and Manufacturing

Executive Summary
Minute Read

Our FDA/Food, Drug & Device Team examines how the Food and Drug Administration is approaching the preclearance and approval process for artificial intelligence (AI) and machine learning (ML) components in products under its purview.

  • Three policy papers in the last year have tackled AI/ML in drugs, drug manufacturing, and medical devices
  • The FDA has discussed a regulatory framework for using AI
  • Companies should remain involved in the ongoing regulatory process to help the FDA design workable solutions

Artificial intelligence (AI) has become the mantra for a new technological revolution, including use in drugs, medical devices, and combination products regulated by the Food and Drug Administration (FDA). On March 15, 2024, the FDA recognized these advances by issuing a new discussion paper to explore relevant considerations on the use of AI and machine learning (ML) in product applications for medical devices. The FDA is continuing to solicit feedback from product designers as it advances its knowledge and modifies “regulatory science” (e.g., application requirements for AI device components). This new discussion paper follows two similar policy documents issued by the FDA on May 10, 2023 related to drugs and drug manufacturing.

On April 11, 2024, in testimony before the House Committee on Oversight and Accountability, FDA Commissioner Robert Califf stated that the FDA efforts are focused on maintaining the United States as “a leader in AI/ML medical devices, including by ensuring that the Agency’s approaches keep the United States ahead of its global competitors and so AI is developed and deployed responsibly in healthcare.” The commissioner’s comments signaled that the FDA is actively meeting with industry groups to design a regulatory pathway that is streamlined and transparent and alerts sponsors to the expected content of preclearance and approval applications containing AI/ML components and related software algorithms.

CDER Digital Health Center of Excellence

On September 22, 2020, the FDA created the Digital Health Center of Excellence within its Center for Devices and Radiologic Health to lead policy review and industry direction. This reorganization was modeled after the creation of the Oncology Center of Excellence within the FDA Center for Drug Evaluation and Research (CDER). It was intended to signal FDA recognition and prioritization of advancing digital technologies through: (1) connecting developers and building strategic partnerships; (2) sharing a base of knowledge to assist in design and review; and (3) reducing regulatory burdens and bottlenecks so this technology is available to U.S. patients first. In September 2023, an FDA public workshop was held to discuss a regulatory framework for use of AI in drug manufacturing.

Presidential Executive Order 14110

A government-wide effort to secure and develop AI was accelerated by Executive Order 14110 issued on October 30, 2023. In that detailed order, President Biden required government departments and agencies, including the Department of Health and Human Services (HHS) and FDA, “[t]o help ensure the safe, responsible deployment and use of AI in the healthcare, public-health, and human-services sectors.” HHS was ordered to establish an AI Task Force to develop a strategic plan including actions to deploy use of AI and AI-enabled technologies in the health sector (including research and discovery, drug and device safety, health care delivery and financing, and public health).

What Has Come Out of All This Effort?

First, the FDA has been quick to acknowledge that it has a base of knowledge acquired by approval, authorization, or clearance of over 700 AI/ML devices, including:

  • Cardiac ultrasound software that uses AI to guide the user.
  • An AI-based device to assist clinicians in detecting lesions (polyps or suspected tumors) in the colon in real time during a colonoscopy.
  • A diagnostic aid for autism spectrum disorder.
  • An AI-based device to detect greater than a mild level of diabetic retinopathy in adult diabetics.

According to CDER Director Patrizia Cavazzoni, in 2021 alone, more than 100 drug and biologic applications submitted to the FDA included AI/ML components.

The three FDA discussion papers contain the same basic information:

  • FDA product centers are working together to identify AI uses and review methodologies. They seek to cultivate collaborative partnerships to examine transparency, explainability (through guidance and otherwise), governance, bias, cybersecurity, and quality assurance.
  • AI regulatory policy should provide predictability to sponsors and encourage use of AI, detect potential knowledge gaps, and highlight opportunities.
  • The FDA is developing a methodology to test and evaluate AI algorithms.
  • It is building upon existing initiatives and prior product review experience.
  • New guidance will be issued on topics including change control plans for AI-enabled device software functions, life-cycle management, premarket submission recommendations, and FDA use of its own AI for application review.
  • Evaluation of long-term safety requirements and use of real-world performance monitoring.
  • Requirements for training and testing of AI models.
  • Framework and strategy for quality assurance.
  • Global harmonization of AI requirements for review, use, and oversight.

What Happens Next?

  • AI/ML policies remain a “work in process” at the FDA.
  • The FDA continues to reach out, review external comments, and monitor developments while reviewing applications for products incorporating AI/ML on a case-by-case basis using existing methodologies.
  • Like gene therapies at CBER, FDA product centers depend on their ad hoc experience to develop general guidance documents based on their internal experience conducting AI/ML product reviews.

Recommendations for Companies in AI/ML Space

  • Review and comment on draft AI/ML guidance documents that should be rolling out shortly on topics including methods to test, validate, and prepare AI/ML marketing submissions; recommendations for predetermined change control plans for AI-enabled device software functions; and life-cycle management considerations.
  • Request interactive pre-submission meetings with FDA product centers to educate the FDA on AI technologies and seek FDA listing of clinical and product performance data requirements.
  • Monitor the dockets opened by the FDA to accept comments and information from industry and outside constituencies.
  • Participate in upcoming industry workshops, Part 15 hearings, advisory committee meetings, and other opportunities to engage with FDA decision-makers.

Please contact us if you require assistance in monitoring, organizing meetings, or drafting and submitting comments relevant to products utilizing AI/ML technology capabilities.

You can subscribe to future advisories and other Alston & Bird publications by completing our publications subscription form. If you have any questions, or would like additional information, please contact one of the attorneys on our FDA/Food, Drug & Device Team.

Meet the Authors
Media Contact
Alex Wolfe
Communications Director

This website uses cookies to improve functionality and performance. For more information, see our Privacy Statement. Additional details for California consumers can be found here.