The FDA's recently established Digital Health Advisory Committee (DHAC) took a significant step last week with its first meeting. This committee is dedicated to ensuring the agency closely monitors medical devices equipped with GenAI throughout their entire life cycles. A 30-page document provided to committee members ahead of their November inaugural meeting sheds light on their concerns and considerations. Let's delve into the key aspects discussed.

Why a Total Product Life Cycle (TPLC) Strategy is Critical

The FDA has long been committed to a TPLC approach, and this becomes even more relevant for medical devices incorporating rapidly evolving GenAI technologies. As these devices are intended to iterate more frequently during their use life, a TPLC approach helps manage them effectively. "A TPLC approach is likely to remain important to the management of future, safe and effective GenAI-enabled medical devices." This ensures that safety and effectiveness are maintained across the device's lifespan.

How the TPLC Approach Relates to the AI Lifecycle Template

For manufacturers of GenAI-enabled devices, considering the FDA's AI Lifecycle is an important way to manage their products throughout the TPLC. "Additionally, the AI Lifecycle can be used as a helpful model to identify where new techniques, approaches or standards may be needed to assure adequate management of these new technologies across the TPLC." It provides a framework for addressing the unique challenges posed by GenAI.

Defining 'GenAI'

GenAI refers to the class of AI models that mimic input data structures and characteristics to generate synthetic content. This includes various forms such as images, videos, audio, text, and digital content. "GenAI models can analyze input data and produce contextually appropriate outputs that may not have been explicitly seen in its training data." Understanding this definition is crucial for proper regulation and oversight.

GenAI vs. Traditional AI/Machine Learning

Like other AI/ML models, GenAI models are developed on large datasets that human developers often can't fully understand during development. However, datasets for GenAI model development can be intentionally broad and not initially tailored to a specific task. "In contrast to the datasets used to develop other AI/ML models, datasets for GenAI model development can be intentionally broad and may not be initially tailored to a specific task." This difference impacts how GenAI operates and is regulated.

What Makes GenAI Tricky to Regulate

GenAI's ability to handle diverse, new, and complex tasks can lead to uncertainty about the limits of a device's output. "When insufficiently controlled, this uncertainty can translate to difficulty in confirming the bounds of a device's intended use, which can introduce challenges to FDA's regulation of GenAI-enabled devices." Regulators need to carefully navigate these uncertainties to ensure proper oversight.

Why Foundation Models Matter

Foundation models are trained on a wide range of data and can be applied to multiple AI applications. If a manufacturer uses a foundation model in a medical device with a specific intended use, it becomes the focus of FDA's device regulatory oversight. "If a manufacturer uses a foundation model or other GenAI tool as part of a product with a specific intended use that meets the definition of a medical device, the product that leverages the foundation model may be the focus of FDA's device regulatory oversight."

How to Avoid FDA Rejection

Manufacturers and developers should consider whether a GenAI implementation is beneficial to public health. If it could provide erroneous or false content, it may not be the best technology for a specific intended use. "It is helpful for manufacturers and developers to consider when GenAI may or may not be the best technology for a specific intended use." Additionally, performance evaluation methodologies will depend on the specific intended use and design of the GenAI-enabled device. "As with all devices, the totality of evidence—which may include premarket and postmarket evidence—can support reasonable assurance of safety and effectiveness of these devices across the TPLC."Read the full report.