Today in health, it, we take a look at the AI speed bumps. Are there any what's slowing us down? What will slow us down? That's what we're going to take a look at. My name is bill Russell. I'm a former CIO for a 16 hospital system. And creator this week health set of channels and events dedicated to transform healthcare. One connection at a time. We want to thank our show sponsors who are investing in developing the next generation of health leaders.

Short test artist side parlay it's certified health, notable and service. Now check them out at this week. health.com/today.

Three quick things for you. Don't forget. Hey, this week health.com/news. Curated news, just for you. Number two. We are continuing our partnership with Alex's lemonade stand and we have a goal to raise an additional a hundred thousand dollars this year. Love to have you be a part of it. Hit our website top right-hand column.

Click on that to give today. And finally share this podcast with a friend or colleague uses his foundation for daily or weekly discussions on topics that are relevant to you and the industry. You can subscribe wherever you listen to podcasts. Alright story. MSNBC companies want to move fast with AI adoption, but see plenty of speed bumps. And they have their key points at the top.

And then they go into this, the article, let me get my glasses. These are pretty small. Let's see barriers to generative AI adoption of bound keeping companies from moving as fast as they'd like among them. Cybersecurity, threats, talent, shortages, and regulatory delays. Staying current with AI developments is an ongoing process, not a one and done event. Those are their key points.

Let's go down to the article. Obviously. Everybody's looking to take advantage of the latest artificial intelligence tools. And that is the gist of the story. A survey of 120 us senior AI machine learning decision-makers conducted in late 2023 by research and media firm, Foundry and technology consulting firm Sears, S E a R C E. Showed that less than 40% of organizations have successfully deployed an AI project. Less than 40% of organizations.

I wonder. Wow. I'd love to do that survey right now. I'm going to put that into my 2 29 project questions. How many of you have successfully deployed an AI project? So if you're planning to come to one of our 2 29 projects, Meetings be ready for that question. That's a Ford, only 40%. It's early 20, 24,

yeah. There. Let's take a look at the reasons among the biggest barriers to adoption the rise in cybersecurity threats. The Foundry Sears study should the 58% of respondents said data security is a leading barrier. To AI adoption. Wow. I'm not sure that's the biggest barrier. I'd be more concerned about the lawsuits, quite frankly, about the data.

I'd be more concerned about the rework. If I implement something and one of those lawsuits happens to take hold and screws up some of the models that have been created. That would be my concern Cybersecurity. I guess it depends on what type of project or how you're thinking about using it. There's an awful lot of protections against that.

Anyway. Let's see. There's lack of understanding. About the security vulnerabilities of AI applications that Jake Williamson, a faculty member at cybersecurity research firm. Ian's research AI apps, especially those with using large language models, bring in an entirely new set of vulnerabilities that are poorly understood by most applicants, application developers and security testers.

William said until there's a better understanding of these issues and better tools to help with auditing and defense. Some CISOs are warning that additional risks might not be warranted.

Okay. I'm not hearing that as much, but I will. I'll definitely look into that. The most productive steps companies can take is to get an educated about how AI works.

Yes, of course I'll move on because there's a stuff we all know. And I return on investment. Another barrier is unclear. Use cases for AI. I think this is a bigger one. Actually many businesses are not thinking about which organizational use cases will bring them the biggest return on investment. And I think we in. In healthcare, you'll hear this. Vernacular, if we need to start with the problem I will take it one step further and say, we need to start with the problems we need to solve.

Like the biggest problems, the problems that were give us the biggest return. Maybe not the biggest problems, but the problems that will give us the biggest returns. And we used to always do this with our projects. This was part of our governance process. It was part of our prioritization process.

We had to do an ROI model. On every project now. As I was cynical on some of these ROI models. Cause we had people all over the organization doing these ROI models and the ones we did within it were very detailed. And we had, we had. In fairness, we had a couple of finance people on our staff who were putting these things together.

And so they were and they were, MBA students from Ivy league schools. They were really good at putting these models together. And then I would see some of these others. And it looks like I, in comparison it would be like crayons and typewritten pages.

It was so anyway, regardless. We need to be looking at the, on the return on these projects and and then evaluating the return as we move forward. Because keep in mind, this is the start of these AI projects. Not for health. We've been doing AI projects for a while, but what I will say is this is a start of an age where we're going to see AI. Permeate every aspect of healthcare. And so we need to get good at these types of projects where we have a a thesis.

We have a projected return. We do a test of that thesis. We then evaluate that thesis. We then scale that thesis. And then we evaluate once again, and you won't want to get into that, that mode, that mechanism so that you can start to churn those through in a lot tighter circles. You're doing more and more projects at the same time. With verifiable and and solid results.

That's gonna be the way you're going to want to move forward with these things. And so yes, we need to identify. The highest returning. Finding the right balance of both complexity and impact is critical on how AI will be adopted across the organization. And no complexity and impact. That four quadrant thing is Always always a valuable. Thing to do and put all your projects on that, complexity easy to incredibly difficult, and then returned from low to high. And then you want to stay in as many easy and high return projects as you can, until you've cleared all those out.

And then you start working towards the more difficult, but high return projects. Again, that's a great graph to to work from. Many organizations want to use AI in applications, but don't know where they can get value from it. Williams said today, we're seeing a bit of a gold rush feeling and don't get left behind

without any real thought as to the applicability. To a specific use case. He said, this reminds me a lot of the early blockchain days. Oh, that's interesting. I wouldn't equate this blockchain, however I could see I see what he's saying. And actually I would not be worried that the don't be left behind thing. It's not something we should be worried about.

I would be worried about. Competitors getting a distinct advantage over us. And so I would be looking at. I'd be scouring articles and looking for ways that people are utilizing the technology that I would identify as potential game changers into how we interact with our consumers, how we deliver on our results.

Be those clinical results. Be those financial results. Be those operational results, whatever it happens to be. Let's see. Finally, AI adoption might be slowed by regulatory policies and compliance efforts with AI adoption being in its early days. Regularly, regulators are still evaluating its implications. Most government and regulatory bodies are still in the early days of formulating the guard rails that will define how AI is more broadly adopted across companies. We're not going to be writing AI models.

So I. It's we are going to have to adopt the right partner who is going to be able to handle these guardrails. And handle the regulatory environment. For us I know a very few. Organizations that are even playing around with the open source models. I'd like to see more of us playing around with the open source models, but only with regard to a promise of some kind of return on the other side.

I wish there was more money in the startup space at this point, because I think there's a lot of opportunity and that's what we're going to miss by the way, in the startup space, where the money has receded, the tide has gone out as it were not as much risk being taken. By investors and whatnot.

We're going to miss a lot of those startups who were given a couple of million dollars. They were going to get a bunch of programmers together and do some proof of concepts essentially is what that was for us. It was really a nice environment that we had there had going there for quite a number of years. I think that will return, but right now it is all about your ability to to cashflow as an organization.

And so a lot of those risks have gone away. Again, on the regulatory side. The one last thing I will say is these lawsuits do have me concerned a little bit. In terms of how much of the data are they going to have to strip out of these models that they train them with? And specifically the chat GPT model I'm most concerned about?

Where did it get its training from? There's a lot of a lot of studies that have gone on, Hey, we've been able to identify where it's got his training from and. They don't really have access to that data. I don't know how that's going to impact that. I'm not sure that would slow me down, but it would slow me down. From doing a significant implementations that are going to be ingrained into our processes until I understood that a little bit more.

And hey, you know what, the lawsuits and whatnot they're having their impact that they desire to have, which is they're slowing people down. People like myself are looking at it, going. Do we really want to go into full scale deployment of that? When we could potentially have a bunch of rework. So anyway, companies want to move fast with AI adoption, but see plenty of speed lumps. And I understand where they're coming from and I understand their concerns. I still think we need to move.

We need to be doing a lot of pilots, a lot of tests I'm seeing them done. I've seen people report out on them. They are learning a ton. They're learning on. Adoption, they're learning on what aspects these large language models are good at what aspects they're not good at how to integrate them into the workflow, how to integrate them with other machine learning models that we've had for years. How to create transparency in those models.

We're learning a ton and that in and of itself is worth a lot because at some point you are going to have to move fast and having that learning behind you is going to be well worth it. All right. That's all for today. Don't forget, share this podcast with a friend or colleague, keep the conversation going. We want to thank our sponsors who are investing in our mission to develop the next generation of health leaders. Short test artist side parlance, certified health, notable and 📍 service.

Now check them out at this week. health.com/today. Thanks for listening. That's all for now.