Responsible Generative AI in Healthcare and What Leaders Need to Know for 2026 - with Rhett Alden of Elsevier Health

Daniel Faggella (0:12)

Welcome everyone to the Emerge AI in Business Podcast. Today's guest is Rhett Alden, Chief Technology Officer at Elsevier Health Markets. Rhett discusses why transparency and claim level validation, tracing every AI-generated ascitation back to peer-reviewed sources, are essential to earning clinician trust in generative AI. He also explains how healthcare institutions are standing up cross-functional governance boards modeled on clinical trial review structures to evaluate risk and oversight as regulation continues to lag behind the technology. A quick note for our audience that the views and opinions expressed by Rhett on today's program are his own, and do not reflect those of Elsevier Health Markets or his leadership. Getting in front of enterprise AI buyers is not about impressions, it's about trust. At Emerge, we help AI vendors engage decision-makers through research-driven content and conversations that actually matter in the buying process. To learn the exact strategies we use to help leading AI brands and startups connect with their ideal enterprise AI buyers, visit emerge.com/ad1. That's emerj.com/ad1.

Now the conversation with Rhett.

Rhett Alden is back in our studio today. Rhett, welcome and thank you for coming back for a third time.

Rhett Alden (1:44)

Thank you. I guess I really love this. Thank you enough, so glad to be here.

Daniel Faggella (1:48)

It goes both ways. In our last discussion, we got into clinicians that might be struggling to trust AI and how they should be using it to make their lives easier.

And today we get to chat about the not-so-glamorous part, the guardrails, the governance, regulations that we should be focusing on. And focusing on responsible generative AI in healthcare, what is blocking generative AI from your clinical operations?

Rhett Alden (2:15)

I would say the biggest challenge for clinicians or for health institutions is the fear of black box algorithms. They're used to things being validated scientifically. They're used to predictability, very used to precision.

And the challenge with generative AI is, how often does it hallucinate?

Are the findings grounded? How sycophantic is the model?

These are big issues, and they're real ones when it comes to things like scientific information. So this is probably the number one barrier for clinicians in general. So that reticence they have is really driven or juxtaposed by the need for quality and patient efficacy, meaning I need to treat the patient properly to minimize harm. So these two tensions are there. A, it's a black box model. B, I need high quality. So transparency is something that needs to be really driven home in order for these tools to be adopted in any clinical operation. Transparency in how the responses arrive that, and also transparency in where the information comes from. These two things are really critical. I think this is why there's such a gap right now between what I would call consumer-based tools and tools built for professionals. Because the professional tools, in our case, is one example. We have a very robust validation framework for CKAI that actually publish results around number of errors, risk of hallucinations, et cetera, to provide that kind of grounding for clinicians. That's really important in this kind of environment.

Daniel Faggella (4:16)

Absolutely. Do we also see the lack of certainty in terms of compliance and regulations as a blocker? I know that in the US, we're currently seeing state-level regulations being implemented. But in general, do we also see that as a blocker? And the fact that we don't really have clarity on regulations and compliance when it comes to AI.

Rhett Alden (4:42)

AI has moved so quickly. Regulations have had a hard time keeping up.

Just recently, there's a new ISO standard, 42001, around AI, that's mostly focused on security issues, data leakage or prompt manipulation, those kinds of things. And that's now showing up in proposals and requests for different systems. But the reality is that I think that people are struggling around what it means to have a quality system. The FDA in the US is not regulating AI at this point.

States are trying to regulate it, but again, it's more patient privacy or privacy issues than it is on anything else.

So we, at Elsevier, we're trying to establish our own principles around responsible AI, trust factors, trust cards, different aspects to give that level of credibility, and not really waiting for regulatory bodies to step in. Because I think it's really needed, but right now it's self-regulation for the most part.

Daniel Faggella (5:56)

And I think that's something that it's not necessarily scary. We know what's right and what's wrong, but it does create a sense of uncertainty for clinicians when we're not exactly sure what should be and what can happen tomorrow. Will we see new legislation? But like you said, the legislation and regulations can't keep up with the speed that we see AI evolving in the industry.