For the medtech industry, the presence of artificial intelligence (AI)-based software and functions within devices, from MRI scanners to ablation catheters, has become ubiquitous.
The US Food and Drug Administration’s (FDA) available pathways for a product’s entry onto the US market have remained largely unchanged for decades. To account for AI-based products, AI-specific guidance and adaptations have until now been grafted onto existing evaluative routes to market such as the FDA’s 510(k) and de novo pathways. Yet with Generative AI (GenAI) entering the picture, this is now likely to change. GenAI is an advancement on more basic forms of AI given its higher reasoning capacity to make decisions independently. With this feature, examination into regulatory frameworks has begun.
Go deeper with GlobalData
Discover B2B Marketing That Performs
Combine business intelligence and editorial excellence to reach engaged professionals across 36 leading media platforms.
In August, the FDA released a 30-page discussion paper outlining a more formalised approach to the regulation of GenAI-based medical devices. Highlighting that GenAI-enabled medical devices hold transformative promise for patient care and the broader health ecosystem, the paper also noted that such devices hold the potential to introduce unique risks compared to existing technology.
The agency’s paper outlines considerations around pre- and post-market evaluation – all with a view towards crafting a regulatory approach to GenAI-based medical devices that “keeps pace with the rapid innovation of digital health technologies”, CDRH director Michelle Tarver said following the paper’s release.
The stakeholder comment period, which closes on 19 October, follows the release of a similar document by the Medicines and Healthcare products Regulatory Agency (MHRA) in June 2026 around its approach to AI regulation in healthcare.
Both documents arrive amid an increase in scrutiny towards the threat that AI may pose. Ex-Anthropic researchers, for example, have warned about irresponsible development of AI, prompting reviews from tech companies.
To learn more about the FDA’s regulatory proposal for GenAI-based medical devices, Medical Device Network spoke with Dennis Gucciardo, partner at law firm Morgan Lewis, about the key motivating factors behind the discussion paper, among other topics.
This interview is an abridged version of a longer conversation with Gucciardo. You can listen to the full conversation here.
Ross Law (RL): What is your response to the FDA’s proposal paper?
Dennis Gucciardo (DG): I think it’s a great first step in figuring out how to best regulate these types of products. The reality is that they are coming to market, whether the FDA likes it or not, and so they’ve got to start thinking about how to best be effective.
The FDA has been regulating software since the early ’90s, but that regulatory model was built on the premise of taking a picture in time, assessing a product, making a determination, and once satisfied, letting it go. And that’s just not comparable to the speed the speed of AI, so what this paper does is continue to push and qualify how the FDA is going to regulate products that benefit from fast development, but without getting in the way.
RL: What do you view as the key purposes of the FDA’s proposal?
DG: Fundamentally, I think the FDA is trying to figure out how they can regulate this technology and continue to protect public health, but also allow the technology to unleash itself.
Looking at the guidance document, it’s clear that the FDA, as ever, is seeking to apply a risk-based approach to GenAI regulation, with the understanding that the power of GenAI is in its decision-making capabilities, but that’s also where the technology holds potential risk.
The power of AI is in its ability to change rapidly. And if you have to keep coming back to the FDA with a submission, which is effectively requesting a change, the technology or software is likely to get slowed down, and companies with such technology may end up losing some value there.
As is effectively being acknowledged in the FDA’s paper, it will need to be comfortable letting itself off and letting some of these products into the marketplace, while also letting the marketplace dictate how it operates and then trusting the company that it will modify the software. That’s really the post-market performance monitoring piece. In the past, the FDA talked about this with the pre-certification programme, and we are seeing it in de novo applications for AI products where there is a post-market component to it. At present, if you’re a medical device, one of the general controls of the Federal Food, Drug, and Cosmetic Act (FDCA) is the need to apply the Quality Management System (QMS) regulation, which includes a post-market component for taking the information, putting it into the risk continuum and making changes to the software.
Right now, that’s all being done through a company’s compliance that’s only checked through inspections. But maybe that’s an area that needs strengthening. And I think with that post-monitoring piece, as reflected in the proposal, the FDA recognises that they’ve got to get comfortable around more steadfastly determining what that should be, and how the regulatory posture should look, so that they don’t have to be involved in every software iteration.
RL: Looking ahead, how do you feel the proposal could impact your client counsel?
DG: One factor is that it gives clients heightened confidence because there’s this position paper out there. A lack of information makes them think that the hurdles are higher because the FDA is not thinking about this.
The fact that the FDA put this paper out suggests they have been talking with people throughout the medical device industry, and that they feel they need to advance their regulatory criteria. And I think that gives companies confidence to develop in this space. In my view, the other thing it does is that it allows us to have arguments ready for when we go to the FDA in terms of how they should think about the product, whether or not they should even regulate the product, and whether or not it’s within the public health to do that. These are all things that the paper provides as we think about working with clients operating in this space.
