On January 6, 2025, the Food and Drug Administration (FDA) released draft guidance aimed at regulating the marketing submissions and lifecycle management of artificial intelligence-enabled devices.

The guidance highlights the necessity for postmarket performance monitoring, as the effectiveness of AI devices may change over time due to variations in input data or patient demographics.

Xiaolan Wang from IMO Health noted that regulatory clarity remains a major barrier to AI adoption in the pharmaceutical industry, highlighting the need for clear guidelines.

This guidance is a response to the rapid growth of AI technology and seeks to ensure the safety and efficacy of products under FDA regulation.

In addition to the device guidance, the FDA also published recommendations regarding the use of AI in drug and biological product development, reinforcing its commitment to transparency and safety in healthcare.

The 20-page document emphasizes the importance of transparency in AI models, urging developers to disclose how AI is utilized, including the types of models and datasets involved.

The FDA has authorized over 1,000 AI-enabled devices through established premarket pathways, indicating the unique considerations these technologies require.

The FDA's efforts arrive shortly after the European Medicines Agency finalized a paper on the role of AI in the drug product lifecycle, reflecting a global trend towards regulatory clarity in AI applications.

Comments on the draft guidance are being solicited from the public until April 7, 2025, with a webinar scheduled for February 18, 2025, to discuss the details further.

The draft guidance is the first comprehensive set of recommendations tailored specifically for AI-enabled devices, addressing their design, development, maintenance, and documentation.

Manufacturers are advised to ensure diversity in their validation datasets to minimize bias, which can occur from the underrepresentation of certain demographics.

This initiative follows a significant increase in drug submissions incorporating AI, with over 500 submissions recorded by October 2024, primarily in fields like oncology and neurology.