We are buyers of AI systems, I thought it might be good to define AI.
Today in health, it we're going to try to define artificial intelligence. My name is bill Russell. I'm a former CIO for a 16 hospital system and creative this week health. Instead of channels dedicated to keeping health it staff current and engaged. We want to thank our show sponsors who are investing in developing the next generation of health leaders. Short test artist, site, parlance, and service. Now check them out at this week. health.com/today.
Having a child with cancer is one of the most painful and difficult situations. , family can face in 2023 to celebrate five years at this week, health. We are working to give back. We will be partnering with Alex's lemonade. Stand all year long. We have a goal to raise $50,000 from our community. And we are excited to be up over $45,000.
This year. Already. We asked you to join us at our website in the top right-hand column, you're going to see a logo for the lemonade. Stand, click on that to give today. We believe in the generosity of our community. And we thank you in advance. All right. I had a back and forth yesterday on LinkedIn.
And. It really has become apparent to me that we have not defined artificial intelligence well enough yet. It's interesting because we have, , a fair amount of people saying, Hey, we're buying artificial intelligence. For our health system, we're placing artificial intelligence in place. And , we may not understand exactly what it is. And so I pulled up a couple of things. I have a couple of distinctions and this article has.
, four things in it, which I thought were really good, which is why I pulled this article. And it's a tech target enterprise AI. And it's what is artificial intelligence? Let me give you a couple of excerpts. I will talk about it towards the end.
Artificial intelligence is the simulation of human intelligence processes. By machines, especially computer systems. Specific applications of AI include expert systems, natural language processing, speech recognition, and machine vision. And I think that's where we get confused. By the way, I think the first sentence really captures one of the main distinctions, which is simulation of human intelligence. And the second is where we get confused.
, we, we lump everything. That's NLP and speak speech recognition or machine vision. Into the category. Of artificial intelligence and that may not be the case. Okay. How does AI work? As the hype around AI has accelerated vendors have been scrambling to promote how their products. And services use it often.
What they refer to as AI is simply a component of the technology such as machine learning. AI requires a foundation of specialized hardware and software for writing and training machine learning algorithms. No single programming language is synonymous with AI, but Python, R Java. C plus plus, and Julia.
Have features popular with AI developers in general, AI systems work by ingesting large amounts of labeled. Training data, analyzing that data for correlations and patterns and using these patterns. To make predictions about future states. This is old by the way, because there is now. New ways of training. And I'll talk about that in a second.
And this way, a chat bot that is fed examples of tax can learn to generate lifelike exchanges with people or an image recognition tool can learn to identify. And describe objects in images by reviewing millions of examples. New rapidly improving generative AI techniques can create realistic texts. Images music and other media.
Let me give you a, another learning model. They used to do this. They used to just ch. You'd say all right. The machine needs to understand what a ball is and they have a million pictures of balls. And they'd have a team of people. It likely in another country. , looking at the image and essentially saying that's a ball, that's a ball, that's a ball. But if you put the ball in a different context, like say on the moon, it wouldn't be able to identify the ball.
Now a child, a human child. If you showed it at a ball and then you showed it a ball on the moon, it would just say that's a ball. , but the computer could not make that distinction because it was not learning. It was just essentially. Being taught. This is what the pattern looks like. This is what the pattern looks like. And with reinforcement.
The new learning models are much more sophisticated and really interesting to me. W what you can have as the self-learning models. And as , computers can do things very rapidly. And so take a take computer vision. It's easiest one to explain here. So you take computer vision. And you say, , you take a image of a ball.
And you break it down into quadrants. And the quadrants are, let's say 20 by 20 quadrant or 10 by 10 quadrants. And what the computer does is it has the image. And then what it does is it will blank out certain areas of the image and see if the computer can recreate the image. So let's say it takes three blocks, blank some out, and then it tries to recreate it. Then it goes back to the original image and it looks, and it says, wow, I I've only recreated it at an 80% level.
And then it tries it again with different blocks, different images until it learns to recreate what a ball looks like. , in those things and because it's doing that rapidly, just over and over again. It can learn a lot faster than the old model, which is send a million images to a bank of humans, which we're going to train it.
So we now have self-learning models, much more, rapid, much more accurate. And it goes on and talks about let's see, 1, 2, 3, 4 characteristics. And I like these characteristics. Which is again, why I pulled this article. AI programming focuses on cognitive skills that include the following.
Learning the aspect of AI programming focuses on acquiring data and creating rules. For how to turn it into actionable information, the rules which are called algorithms, provide computing devices. With step-by-step instructions for how to complete a specific task.
Now the interesting thing to me is that all software uses algorithms. And so just because things use algorithms or even sophisticated algorithms do. Not make them AI. And we're going to go on and talk about that.
However AI are learning systems. They learn, they grow, they get better. Number two reasoning. The aspect of AI programming focuses on choosing the right algorithm to reach a desired outcome. This is why sometimes AI can give us different answers . Because it may hear the inputs and look at it and say, I'm going to use this set of algorithms or this set of algorithms much the same way the human brain works.
We put things into context, and then we answer based on the context that we put it in. And sometimes we may put it in the wrong context. And because we put it in the wrong context, we will make the wrong determination. And that gets to the next one, which is self correction. This aspect of AI programming is designed to continually fine tune algorithms and ensure they provide the most accurate results possible. When you think about a.
A human. As we get better at identifying the context of a conversation, the context of a question. , , the surrounding that we're in. We can self-correct we can learn grow reason better based on what we have learned and the same. Thing's true with AI. And then finally creativity. This aspect of AI uses neural networks.
Rules-based systems statistical methods and other AI techniques to generate new images, new texts, new music, and new ideas. This is the new frontier. This is where, , We have these tools now where you can give them text prompts and they are generating, , images for you. They are generating new art. They are generating new music, poetry, you name it, it's generating those kinds of things. So you have these four characteristics and I like them because they are the distinctions learning reasoning.
Self-correction and creativity. So simply. One of the things I talk about quite, , quite often, especially in healthcare. With regard to what is AI? Is AI machines learn. Whereas algorithms do not learn. If you put data through an algorithm, it will spit out the same answer every time. It's one of the things we love about technology, and we love about computers. You ask it the same question. It returns the same answer.
That is not the case with AI. AI will choose a different set of algorithms. It will essentially adjust based on context much the same way. The human brain does. And so, , one of the distinctions I talked to people about is the more reps, the more opportunities that AI has to answer questions, the better it should get it should get smarter.
The first time that identifies a ball. You tell it, Hey, here are five images of a ball. It should then go out and 50% of the time be able to identify a ball and a picture. As you give it reinforcement, as it begins to understand the context, the next time around it should be 70. 80 90, a hundred percent.
AI. Learns AI adjust AI gets better with repetition. AI gets better with more data. And that is one of the primary differences, especially in healthcare for us. Of AI systems. Algorithms, we'll give you the same answer every time they will not deviate. Quite frankly, just run the data through it comes out the other side and it gets processed. And so we have very sophisticated algorithms in healthcare, which people are calling AI.
They are not AI. AI has these , four characteristics. They are learning. They're constantly improving. They have reasoning self-correction and creativity. That's the distinction of AI. Why do I think that's important? Because we are buyers of AI, we should be able to identify it. We should I identify , the false hope.
That players are bringing to us and saying, Hey, this is an AI. System and it can do these things. And the question I was asked is, does it get better with reps? Does it get better with more data? Are we going to be essentially sitting in front of a new system? And if it is. , a system that gets smarter and better with reps. One of the things we have to worry about is it may give you a different answer.
With different contexts and different questions. That's good. In most cases, in some cases we still expect to computer. To spit out the same answer every time. And in healthcare. That may be important. So we need to understand the distinction between let's just call it legacy code, which is algorithms.
And AI algorithms, which actually learn, adjust and grow with time. So anyway, that's all for today. If you know someone that might benefit from our channel, please forward them a note. They can subscribe on our website this week, health.com or wherever you listen to podcasts. Apple, Google overcast, Spotify, Stitcher, while citrus going out of business, but you get the picture.
We want to thank our channels. Sponsors who are investing in our mission to develop the next generation of health leaders, short test 📍 artist, site, parlance, and service. Now check them out at this week. Health. Dot com slash today. Thanks for listening. That's all for now.