Today in health IT, we are discussing Microsoft CEO admits that AI is generating basically no value. The real benchmark is the world is growing at 10%. My name is Sarah Richardson and I'm president of community development at this week health, where we host a set of channels and events dedicated to transforming healthcare.

One connection at a time, and I am joined by Kate Gamble, Managing Editor here at This Week Health. Kate, welcome to the show. Thank you, Sarah. This topic is super interesting because we're once again discussing AI, but specifically we're looking at recent comments from Microsoft CEO Satya Nadella, who suggested that AI isn't generating significant value.

It's a bold statement given the hype, and we're going to break down what it means for healthcare, IT leaders, and the industry as a whole. Yeah, this is a conversation that should concern everyone in healthcare IT and everyone in general, I would say. We are making huge investments in AI for predictive analytics, patient care, cybersecurity, and more.

So if AI isn't delivering tangible ROI, that is a wake up call. It is a wake up call. So we're going to talk about five things from this article and apply them to the healthcare environment. We're also going to talk about what leaders can take back to your teams, your board, and your peers. So first up, the gap between AI expectations and actual outcomes.

Nadella's point is that while AI is incredibly powerful, businesses aren't necessarily seeing direct value from it yet. And this is where it gets tricky for our health IT leaders. We're being pushed to integrate AI into everything. Clinical workflows, diagnostics, cyber. But if the outcomes aren't measurable, Are we just chasing hype?

And we know that is a huge concern with the people we speak with. Especially with so many components of AI being embedded into EMRs, you have to decide which ones you're going to utilize or otherwise. And I think the question that healthcare leaders really need to ask is, are we implementing AI because it solves a real problem or because it's a trend?

That decision making between which pieces are turned on and off, if it's embedded or integrated as a point solution, which we talked about the day before. Is something that needs to be top of mind throughout this current journey. Yeah. And I would say that it isn't because it's a trend, but because there's a lot of pressure.

There's pressure from inside the organization, outside the organization that why aren't we doing this? So it is a real problem. And that brings us to the second highlight, which is that Nadella pointed out that while AI has massive potential, we're still figuring out how to translate it into increased productivity.

That's where it's critical for health care. So look at AI assisted clinical decision making. We're throwing models at massive data sets, but are we really reducing the clinician burden? And how are we improving patient outcomes with that information? Yeah. And are we adding complexity? Many AI tools require additional validation, oversight and governance, especially when you're dealing with patient data and compliance.

If AI isn't reducing time and effort, then is it really productive? Which brings us to our third point, which is the cost versus the ROI, because it certainly is not inexpensive. Enterprise level AI solutions require massive investments in infrastructure, training, I will pause in that moment though, Kate, to say, if your environment needs a facelift or a refresh anyway, this could be one of the catalysts and be careful about it being the catalyst, because this is an iterative time for us to figure out how we're going to best utilize this in our environments.

So you don't want to over. The ability to do this, and it gets really tricky in the cloud or in different environments really fast in terms of how you're going to use different components for storage and bursting and information, et cetera. So have all of those things be something that your CTO and your CISO are super comfortable with.

And if they're not crowdsource your pool of talent with your partners. So if I'm a CIO and making some of these decisions, I'm going to bring my core partners together for this conversation. So we are equally accountable in how we are thinking about delivering this to the best value and cost within the healthcare system.

Yeah, that's a great perspective. And of course, when we're talking about healthcare, the stakes are higher. It's not like retail or entertainment. We're dealing with patient lives. So if AI doesn't deliver tangible results. Are we spending millions on an illusion of innovation? That sounds like a Lady Gaga song.

Illusion of innovation. Her new album just came out, so I've been on a little bit of a trend with her. But healthcare IT leaders need to seriously ask. What is the return on investment? Are AI powered chatbots actually reducing administrative burden? Are predictive analytics models improving patient flow in hospitals?

If the answer is unclear, or the data is not yet well formed enough to make the next set of decisions, it's time to rethink the strategy or rethink some of the use cases that are being implemented.

And that gets into another point, which is around security. One of the biggest challenges AI presents in healthcare is security. More AI means more data being processed, more data being stored, and more data being shared. And with cyber threats growing, we have to ask, are we exposing more vulnerabilities?

AI generated data models can be manipulated. Bad actors can exploit AI decision making. Plus ethical concerns, you and I often talk about AI bias in healthcare that can lead to misdiagnoses, disparities in treatment recommendations, and even ethical breaches in patient privacy. So where are your models being stored?

What is the source of information? How do you bring external views or data into the modeling? What is the governance around all of those components? Yeah, so leaders need to ensure that AI deployments aren't just about speed and efficiency, but also about security and ethical governance. Because we needed another element to governance, Kate.

We needed the ethics component, too, which has always been there, yet governance, which is already really difficult, now has additional elements being layered in. So I'm just hopeful to all of our listeners and to our community that governance is something that still stays at the forefront, especially as the speed by which some of these new decision making, So with all of these concerns, we cannot overlook the AI is making a measurable impact in some areas of health care.

Yeah, there is a reason for the hype, right? Looking at some of the key places where AI is delivering real, tangible value, one is radiology and imaging. AI helps radiologists detect abnormalities like tumors, fractures, and lung diseases faster and more accurately. And that's a huge draw. And then, as far as predictive analytics and early disease detection, AI models can help identify high risk patients for conditions like sepsis and heart disease, allowing earlier interventions.

There's also benefits in terms of personalized medicine and drug discovery, and administrative efficiency and workflow automation. With that, you're talking about chatbots, NLP tools, and automation, which can reduce documentation burdens for clinicians. And of course, we've seen some benefits in cybersecurity and fraud prevention.

AI can enhance threat detection, protecting patient data from cyber attacks. And the list goes on. We have robotic surgery and AI assisted procedures. And of course, remote patient monitoring and virtual health assistance, which is one of the ones that we probably focus on most because we have seen those results, that AI powered wearables and virtual assistants can help monitor chronic disease patients in real time.

This is where healthcare IT leaders can focus tangible AI benefits. These are the ones that align with real clinical and operational needs. And if you're, you have a solution searching for a problem, sometimes we hear that, and you've got some of these things that are already embedded. Here's a whole list or a whole space of things where you can try them out without as much risk because it's already embedded in your organization.

And really consider, are these the types of tools that help us deliver better?

And so if we're looking at this on the whole, how do we make AI work for us? It's a big question. It is a big question. So let's wrap it up. AI and healthcare. Is it valuable or is it overhyped? This has come up in several of our interviews, most recently with Keith Perry of St. Jude's Children's Research Hospital, who talked about the role of AI and what it can play in the exam room of the future. And UConn Health CMIO Dirk Stanley, who's excited about leveraging ambient listening for chart summarizations.

But both organizations are taking a very deliberate approach to adopting AI, which is critical. It is, and that's what you want to see, given all the concerns. So I would say that AI is potentially valuable, but with a lot of caution. So if we continue investing without clear outcomes, we're wasting resources.

so leaders need to set clear metrics for AI success. And need to educate their teams, their boards, their peers. We always about those constituents because these discussions aren't about buzzwords. They're about focusing on actual value risks and ethical considerations. And this is a talk track that the C suite of IT and all the way through the organization needs to be comfortable being able Facilitate, moderate, and educate.

Exactly. For our listeners, take this conversation back to your teams. Challenge the AI narrative in your organization. Ask the hard questions. Are we investing in AI because it works, or because we feel like we have to? And that's it for today's episode. Join us tomorrow when we discuss the impact of recent NIH budget cuts on healthcare.

Kate, thank you for joining me today. And remember to share this podcast with a friend or a colleague. Use it 📍 as a foundation for daily or weekly discussions on the topics that are relevant to you and the industry. They can subscribe wherever you listen to podcasts. Thanks for listening. That's all for now.