181: How AI is Revolutionizing Financial & Healthcare Services

Welcome to The Invested Dads Podcast, simplifying financial topics so that you can take action and make your financial situation better. Helping you to understand the current world of financial planning and investments, here are your hosts, Josh Robb and Austin Wilson.

Austin Wilson:

All right. Hey, hey. Welcome back to The Invested Dads Podcast, the podcast where we take you on a journey to better your financial future. I am Austin Wilson, Research Analyst at Hixon Zuercher Capital Management.

Josh Robb:

And I’m Josh Robb, Director of Wealth Management at Hixon Zuercher Capital Management. Austin, how can people help us with our podcast?

Austin Wilson:

We would love it if you’d subscribe, if you’re not subscribed already. That plus, follow, whatever button that is on your podcast player so that you get new episodes when they drop each and every Thursday. We’re going on 200 before too long.

Josh Robb:

That’s right.

Austin Wilson:

So do that. And if you’d like another way to make sure you get that new episode, go to our website, subscribe to our newsletter, and then every Thursday when that episode drops, boom, email in your inbox has a link to listen. Some nice show notes. It’s really lovely.

Josh Robb:

Yes.

Austin Wilson:

You should probably do that. So today-

Josh Robb:

Now we got to start sounding like robots.’.

Austin Wilson:

We’re going to start sounding like robots-

Josh Robb:

Because.

Austin Wilson:

So I am doing podcast on AI. As a matter of reference-

Josh Robb:

Yes.

Austin Wilson:

Artificial intelligence, ChatGPT, we’re going to talk about that too. These are some buzzwords right now in the markets, social life, school, and everyone kind of is hearing about them all the time. So we’re going to talk about that.

But interestingly enough, we have used ChatGPT to build the foundation for this episode. So with artificial intelligence, you essentially prompt it and it will give you via natural language, which we’re going to talk about, what you ask for. It’s kind of like search, but better because it’s like-

Josh Robb:

It doesn’t show you where the answer is, it gives you an answer.

Austin Wilson:

It gives you an answer. So essentially I said, “Hey, write a podcast outline about artificial intelligence and ChatGPT.” And you know what? It did and it didn’t suck. So we use that outline. We’ve been tweaking and adjusting, but we got this ball rolling-

Josh Robb:

It’s pretty cool.

 

[2:10] – What is Artificial Intelligence (AI)? 

Austin Wilson:

Using AI. It’s pretty cool. Yes. So what is AI?

Josh Robb:

AI, artificial intelligence.

Austin Wilson:

It’s artificial.

Josh Robb:

Yes, meaning not real intelligence.

Austin Wilson:

It’s real, but it’s not from a brain, right?

Josh Robb:

Yes.

Austin Wilson:

Yes. So artificial intelligence refers to the development of computers, and those computers can perform certain tasks. Tasks that usually require human interference, human intelligence, human problem solving, decision making, and all of these things. So there is a bunch of different ways that we’re going to talk about of how that’s done, that recognize patterns, make decisions, and spit out answers. So it’s used all over the place. It can translate things, it can recognize speech, it can actually analyze video and images, and it’s being used in autonomous vehicles, and it’s just very, very popular right now.

Josh Robb:

And this isn’t new. The concept of a program or system learning from past experiences is not new.

Austin Wilson:

Correct.

Josh Robb:

In fact, that’s been the goal, is that if something keeps having an error, it hopefully will learn and stop doing that. And so an easy example I thought of is, we talked about this past, there’s really cool budgeting software out there. And again, this is just a very simplified version of this. But if I say, “Hey, this transaction is categorized as entertainment.” It learns and says, anytime I see this, I’m going to do that.

Austin Wilson:

That is a form of artificial intelligence.

Josh Robb:

So I mean, we’ve been utilizing or trying to utilize this. We’ve just had advancements along the way to enable this to be even more powerful.

Austin Wilson:

Yeah, and that’s what we’re going to get to later is the power and how the power essentially compounds, right?

Josh Robb:

Yes.

Austin Wilson:

We love compound interest, but what about compounding computing power?

Josh Robb:

Ooh.

Austin Wilson:

Ooh, that’s crazy.

Josh Robb:

Ooh, crazy.

 

[3:47] – What is ChatGPT? 

Austin Wilson:

So another buzzword aside from AI is a more specific version of AI, and that is ChatGPT.

Josh Robb:

What’s that stand for?

Austin Wilson:

I don’t even actually know what it stands for.

Josh Robb:

I’ll tell you.

Austin Wilson:

Okay.

Josh Robb:

Because I Googled it. No, I didn’t. But it stands for going to predate terminators because this is the intelligence that Terminators are going to use to take over the world.

Austin Wilson:

Oh, that’s crazy.

Josh Robb:

I think that’s what it means.

Austin Wilson:

So essentially there’s this company called OpenAI, and they had some influence from Elon Musk back in the day. But they were founded on Open Source, meaning anyone can use the code essentially. It’s public, it’s free software, machine learning. So that’s the idea. So there’s this company called Open AI. Which by the way, Microsoft just took a big stake within 2023 because this whole ChatGPT thing took off. I think they own like 40% of it or something like that.

Josh Robb:

Yeah. They said, you know what? Our little paperclip guy from Office needs an upgrade.

Austin Wilson:

Needs an upgrade. Finally.

Josh Robb:

His questions and answers are not helping anybody.

Austin Wilson:

Oh, man.

Josh Robb:

We need somebody to learn.

Austin Wilson:

Remember those days though?

Josh Robb:

That guy was a helpful guy.

Austin Wilson:

I know, I know.

Josh Robb:

He was like, “Hey.”

Austin Wilson:

Very cheesy. But kind of helpful.

Josh Robb:

I mean, that was the nineties, that’s what you could think of. Clip Art and the little guy right there.

Austin Wilson:

I love me some Clip Art. So ChatGPT is a large language model. Meaning it’s a model that searched the internet up through a point in time, and then formed its own thinking of everything that’s on there to form the answers that you prompted to say. So, it’s looking at the way that language is already in the internet and then it is able to search and combine and do all kinds of crazy stuff there. So, it’s owned and operated by OpenAI, and it’s based on GPT-3.5 to currently working on GPT-4 architecture. This is generation 3.5 to 4 architecture. ChatGPT-3.5, which is what a lot of people are using. The one that I ever use is using 3.5, I don’t have access to 4, I guess I’m not good enough. It’s still pretty wild. But it stops through the end of 2021.

Josh Robb:

Because I asked it a question, it did not like my question. Because it revolved around 2022 data.

Austin Wilson:

And the reason that some of these are set at certain points in time, which I’m sure like ChatGPT-3 or 2.5 or whatever, or is even a point earlier, is that to analyze that vast amount of data on the internet prior to that, it takes a lot of time.

Josh Robb:

You need a stop point.

Austin Wilson:

You need a stop point.

Josh Robb:

New data’s always coming in.

Austin Wilson:

And now I’m sure that as AI becomes more involved, we’re going to get to the point where it’s in real time. But we’re not there yet.

Josh Robb:

I know.

Austin Wilson:

So that’s why there has to be a stop point.

Josh Robb:

Now, just thinking through this, all right, you have a computer, you plug it into the wall.

Austin Wilson:

What if it’s a laptop?

Josh Robb:

Unbridled electricity comes into it. Just that’s what’s powering this thing.

Austin Wilson:

Power.

Josh Robb:

Electricity hits these circuit boards and it’s using codes, zeros and ones is how everything’s coded. To then somehow answer any question, I ask it by going to a database of all other zeros and ones, interpreting everything in there, bringing it back and spitting it out in a way that’s-

Austin Wilson:

And very quickly.

Josh Robb:

Yeah. Ah, just crazy. It blows my mind.

Austin Wilson:

If you type in a query, which is what it would be called. If you type in a query for ChatGPT, if it’s a complex one, it starts spitting out an answer in five seconds. And then it types it out just like it’s speaking, like that fast. And it’s formulating it as it goes.

Josh Robb:

Oh man.

Austin Wilson:

So yeah. So like I said, it’s formulated on large amount of text data using deep learning techniques. Meaning it’s learning from all of this data that’s available on the internet. And then as people are prompting it, it’s learning itself even further and tweaking things –

Josh Robb:

Because if they ask a follow-up, it’s probably learning, oh, I probably should have included that information –

Austin Wilson:

And then the next person might get a better answer. So, this allows it to understand and generate what we would think of as pretty natural language. And it really is, if you ask it to say something, you can actually say it. Explain market capitalization to me like I’m five. It will do that. It’s crazy. Or I said do something and say it like a millennial. Because that’s me, I’m a millennial. And it said it like-

Josh Robb:

Use slang.

Austin Wilson:

Yeah, use words that I would understand. I don’t know. It was just so interesting to me.

Josh Robb:

When you did the one, where you asked it to write a song. Oh man.

Austin Wilson:

I literally tested this using history stuff. So, well below 2021. I said, write a song about Rasputin’s role in World War I. So Rasputin was the advisor to the Czar in Russia in World War I and really had a big impact on that. So anyway, it did. I said, do it in the key of G and it gave me chords.

Josh Robb:

That’s crazy. And to me, music is probably the hardest thing for a computer to understand. I mean, it’s an abstract concept. Man.

 

[8:15] – What is the Purpose of AI & Some Positives? 

Austin Wilson:

So that was interesting. The question is why this is a thing?

Josh Robb:

Terminator, I already told you.

Austin Wilson:

Terminator, obviously. But the purpose is to assist users. So you and me using them, generating human-like responses. So ones that we would come up with on our own to their questions, their queries, their tasks, to provide information as needed. So that’s why.

Now, this has certainly an impact on society. Potentially some negatives, as well. So positive, number one, increased efficiency and productivity. It literally can automate tasks and processes. People can actually work on things that are deemed more valuable, use their brains more. Do all the menial stuff with technology. Another one is it could improve healthcare. So as this continues to evolve, disease diagnostic, personalized treatment plans, these can all be possible through artificial intelligence.

Josh Robb:

Well, I could see that because I’m not a doctor, but I assume doctors have to memorize all the different symptoms to diagnosis combination.

Austin Wilson:

You just plug in, he’s got a sneeze, he’s got a fever and a broken knee, and that becomes-

Josh Robb:

Yeah, this is what it is.

Austin Wilson:

Hip dysplasia.

Josh Robb:

Yeah. I don’t know. That’s weird.

Austin Wilson:

That’s a German Shepherd problem.

Josh Robb:

Yeah, probably. So the idea is just that, it helps them narrow it down probably. Here’s the list of four or five things it could possibly be. And then you, okay, I did some more tests. Narrow it down. Ah, that’s crazy.

Austin Wilson:

Another way is that it could enhance safety. So via internet traffic monitoring or webcam monitoring, all these things. Theoretically, artificial intelligence can monitor and detect threats, like maybe bomb threats or shooters or just sketchy looking things that-

Josh Robb:

Interesting.

Austin Wilson:

Make you want to look into it. As well as cybersecurity. So like bots and all kinds of crazy stuff.

Josh Robb:

This is the sixth car that’s crashed into that tree. Maybe we put a sign up.

Austin Wilson:

Maybe just take out the tree. Disaster response could be another one. So I think that there’s some opportunity for enhanced safety. It also could improve customer service. Now, I hate automated.

Josh Robb:

They have to really improve this to get there.

Austin Wilson:

AI, this is a great opportunity to help because I really dislike calling and having to menu the ones and the twos and the threes to get to where you want to get to. And eventually I just say-.

Josh Robb:

Operator-

Austin Wilson:

All the time.

Josh Robb:

Operator, operator, operator.

Austin Wilson:

I just want to speak to a representative.

Josh Robb:

Person, human being.

Austin Wilson:

I do not want to push a one. So anyway, hopefully we can get to the point where it like sense your tone.

Josh Robb:

When they do the audio cues, I just don’t say what they want me to say apparently.

Austin Wilson:

Really?

Josh Robb:

It’s like, please describe what you’re calling for, and I do. And they’re like-

Austin Wilson:

That’s not an option.

Josh Robb:

Please give better details. I’m like, I don’t know how else to tell you what I want. So apparently I need to be better at that. Maybe it’s on my end, it’s a user area.

 

[10:58] – What Does an AI-Driven Future Look Like & Some Negatives? 

Austin Wilson:

Absolutely. Another one is that it could advance science and research because you’ll be able to analyze large amounts of data quicker and more efficiently, which could allow us to have new insights and discoveries. So those are some of the good ideas that can come from this.

But there are obviously some negatives too. Number one, and this is the one that a lot of people are worried about. Because I think I saw a statistic, and you know what? I’m doing this live, I’m pulling it up.

Josh Robb:

He is going to pull it up live.

Austin Wilson:

Pulling it up live.

Josh Robb:

I’ll fill this time in with awkward silence. Ready, set, go.

Austin Wilson:

Oh, here it is.

Josh Robb:

Oh, you found it.

Austin Wilson:

Goldman Sachs estimates that 25% of jobs can at least partially be replaced by AI.

Josh Robb:

All right.

Austin Wilson:

So job displacement is a real threat because AI can essentially automate many jobs, which could cause a lot of job losses. Another one is privacy and security risk because these systems collect and analyze data that you put in, essentially even personal data. So that could be an issue. There could be misuse. You could create deep fakes, you could create all kinds of stuff that’s sketchy. So that’s an option. And then ethical concerns. So as AI becomes more advanced, it essentially can get more and more advanced on its own. And then, who do you hulk accountable for if AI tells someone to do something, who’s responsible? I think that’s a big legal question in the future.

Josh Robb:

That’s probably would be interesting. Yes.

 

[12:06] – Different Subgroups of AI 

Austin Wilson:

Next we should talk about different subtypes of AI because there are different sub-categories of AI. And number one is machine learning. It’s development of algorithms that learn from data and then improve their performance as they get more data over time. So that’s machine learning. Then we have natural language processing, which we kind of hinted at a little bit. And that involves development of algorithms that can analyze, understand, and generate human language. So just like you and I would talk to each other, it can generate language that sounds like a human.

Josh Robb:

I always see that as a kid learning to talk. It starts simple because they can understand what the idea is. But the more they hear and interact with it, the more they could take a sentence and add verbs and adjectives to it to describe more. Instead of saying, “I want that.” They can give more details. “I’m hungry and I want to eat that one thing that you’re holding in your hand.”

Austin Wilson:

Absolutely.

Josh Robb:

And they can draw all that out. And I think that’s like you’re saying there’s, the more language it can experience and then learn and correct from, the more natural it’s going to sound.

Austin Wilson:

Yep. And this gets better all the time.

Another one is computer vision. So think of-

Josh Robb:

Glasses for a computer.

Austin Wilson:

Glasses for computers. I’m thinking of autonomous vehicles. So this is where a camera sends data to a computer to interpret that visual data, scan-

Josh Robb:

Oh no.

Austin Wilson:

Images, videos.

Josh Robb:

You know what I just thought of?

Austin Wilson:

What?

Josh Robb:

All this security checks to say, are you a robot? And it’s like, click every box that has a bus in the picture-

Austin Wilson:

It’s going to do it for you.

Josh Robb:

Oh, no. It’ll be able to interpret that-

Austin Wilson:

Bot everywhere.

Josh Robb:

I see the buses.

Austin Wilson:

Bots everywhere.

Josh Robb:

Oh no, we’re no longer safe.

Austin Wilson:

We’re no longer safe. But an interesting part about this is facial recognition. Facial recognition’s already kind of useful, my iPhone has it right here. It’s so easy to Apple Pay or open up because of that. But that’s artificial intelligence, which is pretty cool. Another one is obviously robotics, that is artificial intelligence.

Josh Robb:

Terminator.

Austin Wilson:

Terminator. So automation as it relates … again, I mentioned autonomous vehicles as part of it, but industrial automation. Think about the manufacturing process where you could have these robots learn to do things more efficiently over time.

Josh Robb:

Faster.

Austin Wilson:

Faster, more efficiently. That should help profitability for manufacturing, speed throughput.

Josh Robb:

Safety.

Austin Wilson:

Safety, all these things, but-

Josh Robb:

There’s some jobs that you’d want.

Austin Wilson:

Yeah. The other issue is the job impact of manufacturing from there. Another one is expert systems. So things you have to have good expertise in. So think about industries such as finance, such as healthcare. We have to know a lot about things to be able to make decisions for clients, for investments, for patients, for their health. Well, we will be able to hopefully get some help in analyzing, just like you said. Maybe you put in a couple categories or things that are checking some boxes and it’ll spit out what’s going on. That could be kind of cool. Or a threat. Who knows?

Reinforcement learning is another one. This is trial and error. So these algorithms can put in a query, and if it sends back an error, it’ll put it in a little bit differently, until it gets it right. And then that’s going to help it, so you just continually getting it better and improve. So think about game playing. You could use AI to learn chess-

Josh Robb:

Chess.

Austin Wilson:

Yeah. And you could have every single possible move on both sides mapped in minutes.

Josh Robb:

And then, however the opponent plays, it knows then the counter move to-

Austin Wilson:

You could actually-

Josh Robb:

Every move, then creates a new-

Austin Wilson:

You could probably figure out the perfect moves, the absolute perfect moves. Statistically speaking. And I’ve seen this through AI, there’s an example of … it looks like a maze. And AI has this way to plug in, there’s this maze built and you send a ball through. And as soon as the ball gets stuck or whatever in this maze, then that counts as a fail. But the more the ball goes through the maze, based on the algorithm, it figures out which way to go. And eventually it finds the perfect way through the maze. That’s AI.

Josh Robb:

Interesting.

Austin Wilson:

It’s trial and error. And deep learning is the other. And this is the development of those neural networks that can learn from data, perform tasks, image recognition, natural language processing. So definitely some subcategories there. But whew, there’s a lot going on in the AI world today, let’s just say that.

So I guess we kind of mentioned it a little, but we should separate machine learning and deep learning, and understand what those are next. So machine learning is the subset of AI that requires development of algorithms that can learn from data and improve performance over time. But deep learning is different. So that’s a type of machine learning that involves development of neural networks with complex algorithms inspired by structure and function of the human brain. So deep learning is essentially trying to mimic how the brain processes, which is pretty incredible. And this can actually probably process things faster, eventually. Taking large amount of data, including speech recognition, natural language processing, and again, autonomous driving is part of that, as well.

Then we can break down how these machines learn. Well, these machines can learn supervised and unsupervised. So supervised learning, well that involves the training on a model of labeled data, where the input data is paired with output data. Really, you’re giving it two options and it’s mapping in between. Unsupervised is different. So that’s unlabeled data, it’s just random.

Josh Robb:

You take a bucket of data and dump it in.

Austin Wilson:

You can get any output data out of it and that is another way that you can essentially try to analyze data without giving it anything it’s looking for.

Josh Robb:

No preconceived notions.

Austin Wilson:

No preconceived notions, yep. And then we can talk about what natural language processing is, and that deals with the interaction of computers and human language. Always developing natural algorithms that analyze, understand, and generate natural language. So the more data input you have, the more it can understand how humans talk and the more it’s really going to spit out something that sounds pretty real.

So let’s go a little bit deeper into ChatGPT. So how does it work? Generative, pre-trained, transformer.

 

[17:53] – How Does ChatGPT Work? 

Josh Robb:

I was way off.

Austin Wilson:

So yeah, how does it generate those human-like responses? Because it does a really good job, as we’ve talked about before.

Josh Robb:

Well, that I think the job market’s so good right now because they actually have a bunch of people somewhere writing responses.

Austin Wilson:

That’s exactly it. Yeah, yeah.

Josh Robb:

It’s all just fake.

Austin Wilson:

So like we’ve mentioned multiple times, it trained on a massive amount of historical internet data. Which allowed it to learn the patterns, the relationships that humans use when they type different words together. So additionally, it uses a language model, and that language model can actually include contextual information. So somehow if you’re typing something about a car, it knows what kind of words to typically go with that or whatever, which is pretty fascinating. So those responses are contextually relevant and actually sound like a human would sound when they were talking there.

There are a number of ways that this can be used in different industries. So businesses can actually use ChatGPT, it’s free for customer service and marketing and sales. So those are a couple ways to do that. You can actually use this technology to build chatbots, which sometimes I also hate. Because usually they’re-

Josh Robb:

They’re getting better.

Austin Wilson:

They are getting better. But usually they’re very, very bad. But this hopefully will lead to improved customer satisfaction, lower response times. So that’s something there. You can also work on personalizing things like marketing campaigns, product recommendations, advertisements. That you can essentially say, “Hey, write an advertisement for this product that will target this audience.” And it’ll kind of do that for you, which is pretty cool.

So you can use it in sales to get leads, as well, which is a cool way to do that. And again, provide the relevant information that is required for that target audience there. Businesses can use this to engage their customers better and to provide better support. Ideally, because you’re taking away some menial tasks, you can increase operational efficiency too, because maybe you’ll have less head count, or the people you do have are going to be extremely efficient, which is really-

Josh Robb:

Yeah, more time to do other things.

Austin Wilson:

Yep. So that’s pretty cool. There are limitations on the software. One of them is the lack of understanding of the context. So the software can essentially generate a response that could be inappropriate. It might not be right.

Josh Robb:

It doesn’t know.

Austin Wilson:

It doesn’t know. It could be inaccurate. Especially when it’s a really ambiguous request or it’s a really complex topic that hasn’t been queried probably a lot. Another one is that this is a software technology, it does not have the ability to reason. It’s not a brain, it’s not a human. And that’s where humans are always going to have the leg up. We have the ability to do that. We are made to do that. Software cannot.

There are some ethical concerns though, Josh.

Josh Robb:

Yes, always.

Austin Wilson:

And this is, we kind of alluded to it from the legal standpoint of where this could go. But again, this could be misused in terms of creating information for harmful purposes. You could probably put in, create a X, Y, Z that leads people to believe this. And it might do that. I don’t know. I don’t do sketchy things on the internet.

Josh Robb:

You don’t?

Austin Wilson:

Nope. Again, privacy concerns is one problem. And then it’s not very transparent how this works. Now it’s OpenAI, it’s Open Source, but the way that some of these are generated, it’s very difficult to explain.

Josh Robb:

It’s so complex.

Austin Wilson:

Even though it’s public, it’s very complicated. There’s definitely some limited transparency and a accountability from that.

 

[21:02] – Dad Joke of the Week 

Josh Robb:

All right, let’s take a quick break. I have a dad joke for you.

Austin Wilson:

Did you generate this on ChatGPT.

Josh Robb:

No, they did not have a good response. So I think they failed on the dad joke section in that it wasn’t funny.

Austin Wilson:

Humans have the leg up on that too.

Josh Robb:

Yes. They don’t understand humor like we do.

Austin Wilson:

Because we’re really funny.

Josh Robb:

We’re talking … I mean, this is on the internet, so we’re going to stick to the internet area. Weight Watchers, they rebranded-

Austin Wilson:

I need that.

Josh Robb:

They are WW. But I noticed their website needs updated because being Weight Watchers, there was a lot of cookies on that website and I just wasn’t sure that was how they wanted their brand to be.

Austin Wilson:

That’s funny. Yep, that’s funny.

Josh Robb:

Cookies.

 

[21:44] – How Will Artificial Intelligence Impact Society & Industries? 

Austin Wilson:

Back to AI. Two more main sections and then we’re going to kind of give our opinions on how AI is working here. So society is going to be impacted by AI in a number of ways. Overall, it does have potential to revolutionize certain areas of life. It actually could improve quality of life, but there are a lot of challenges that we talked about, as well. One thing that we’re going to talk about a little bit is that regulation is huge.

We don’t really have much regulation in the space, and that’s a concern of some people. But automation of jobs and certain tasks is a reality here. Because human intelligence drives how people work. A lot of these skills and tasks are getting to the point where we can generate software to do it. Will it do it as well? I don’t know. But will it do it as well for a certain cost? That is really where the question is. Because while that automated phone dial or that automated chatbot may not be as good as a customer service representative, it costs pennies on the dollar for what you’d actually pay a real person to do it. And that’s why a lot of businesses have gone to it. So if things can be more accurate actually, because it’s all rules based. There’s not a lot of room for fat-fingering things in this, which is a good thing.

On the flip side, it actually could generate new jobs. So in the areas of technology, AI research, software development, data science, that’s going to be a hot area where there’s going to be a lot of jobs kind of booming because of this. Finance and healthcare are a couple areas we talked about where it actually could have some sort of an impact. And we’re saying this as people working in the finance industry. So in finance, this actually could help with fraud detection. So some algorithms can be trained to detect fraudulent transactions, and that could actually red flag them without a human doing it. Another way is risk management. So if you look at maybe investment portfolios, large data sets, you could analyze the data sets to flag certain risks. That’s one option.

Josh Robb:

Now, do you think there was a big thing with roboadvisors? Which was automated investing for people. Which again, we’ve talked about that in other episodes and I don’t have any bad opinions on it. It’s a great tool for people who just need a little help in setting those asset allocations and aren’t sure exactly what to put it in. I think this could really help enhance that, by taking more data and able to have almost a … I don’t want to say conversation. But take in some more data from that person to-

Austin Wilson:

It’s less of a one size fits all. So you essentially putting your age and you’re done on roboadvisors the way they had been built. But maybe now, you could A, put in your age, but B, put in some thoughts on risk tolerance and loss thresholds.

Josh Robb:

What words you use, how you say things that maybe I’ll learn to say, oh, this may be a good fit. So I thought that was a-

Austin Wilson:

That’s a really good idea. So you’re going to create a software for this, Josh.

Josh Robb:

I’m not. Zeros and ones confused me.

Austin Wilson:

Zeros and ones confuse me.

Josh Robb:

I just don’t … I can’t comprehend how those two numbers-

Austin Wilson:

Turn into everything.

Josh Robb:

Turn into everything. Oh my goodness.

Austin Wilson:

Like I’m looking at a screen with letters and stuff on them, and that’s all even numbers, ones and zeros. And the whole screen colors are ones and zeros.

Josh Robb:

I don’t know.

Austin Wilson:

Okay. The other area is healthcare, talked about it a little bit. But AI can be trained to analyze medical images or electronic health records, which is really cool.

Josh Robb:

Here’s your broken bone.

Austin Wilson:

Yeah, exactly.

Josh Robb:

It’s the one sticking out.

Austin Wilson:

It’s the one sticking out. Or other data to assist with diagnostics and treatment planning. So take away some of the work. Drug discovery, you could put in large amounts of data, predict their efficacy, which could accelerate the drug discovery process. Patient monitoring is another one. You could put wearables and sensors on patients and have AI analyze the trends and then flag things when there’s issues.

Josh Robb:

Is that kind of like your Apple Watch?

Austin Wilson:

I used to have one.

Josh Robb:

Through some things that could say, “Hey, you may want to look into this because your heart rate rhythm is weird.”

Austin Wilson:

Blood pressure, heart rhythm. Yep, it could do all of those things. That’s AI.

Josh Robb:

It takes some of those and learn over time.

Austin Wilson:

Essentially, AI when it comes to things like that, it looks for a baseline and then looks for outliers.

Josh Robb:

Anomalies. Yeah.

Austin Wilson:

If you have an outlier, that’s probably a red flag. And if you have more than one, probably, maybe it should be looked at.

Josh Robb:

Everybody has a heartbeat, you don’t. That’s probably not good.

Austin Wilson:

You might want to get that checked out.

Josh Robb:

Yeah, call a doctor.

Austin Wilson:

Oh, your watch is just on the nightstand.

Josh Robb:

Oh, nevermind.

Austin Wilson:

No worries.

Josh Robb:

Forgot.

Austin Wilson:

All right. So next we’re talk about the future. Because this is where things get a little crazy.

Josh Robb:

This is where the Terminator shows up.

Austin Wilson:

This is where the Terminator shows up. There are some improvements because this is still pretty early, really. We’ve had AI, but if you think about Moore’s law and the exponential ease of creating new computing power or technology, we’re kind of in that in AI right now. It’s still relatively early on and it gets better in real time. But what can be improved is algorithms, because as they go on, analyzing more data and more queries, they’re going to become more and more applicable to wider range of applications. Another is better data handling. As you get more data, again, the AI systems are going to be able to handle those more complex and diverse data sets, which is going to make them more efficient at analyzing them all the time.

Josh Robb:

And organizing that data, I think. Because you can’t just scan 40 years worth of internet data and say, “Oh, find the answer.” It’s got to be able to organize that-

Austin Wilson:

Systematically.

Josh Robb:

And know where to group it.

Austin Wilson:

Another one is increased automation. AI, while it’s getting to the point where it can free up humans for certain things, it’s going to continue to automate certain tasks. Which is actually can be levered for a very good or maybe a bad thing. Who knows? And another one is, this is going to become more of a part of our lives, in general. So human machine AI interaction and collaboration is going to be a way that this just becomes … it’ll become part of a working process. Right now, it’s kind of separate.

Josh Robb:

So you think of the older generation and their struggle to program the time on their VCR back in the day. This new thing, as we age and have a harder time with technology, we may have an in between to say, “Hey, help me to understand how to fix this.” Because all they could do is just yell at it and just throw things at it.

Austin Wilson:

I thought you were going to say, this is going to be what you struggle with.

Josh Robb:

Well, it would be.

Austin Wilson:

You can figure out that the-

Josh Robb:

I could program the clock, but-

Austin Wilson:

Yeah, but using ChatGPT, that’s not your thing.

Josh Robb:

But I’m thinking I could just ask ChatGPT to tell me how to do the VCR coding. I now have an in between. An in between to help me communicate with that technology. I don’t understand.

Austin Wilson:

And actually, ChatGPT, it’s kind of replacing certain things like YouTube and Google searching on certain things.

Josh Robb:

Oh, yeah.

Austin Wilson:

Because you can use it as a search.

Josh Robb:

You actually get an answer.

Austin Wilson:

And you get an answer instead of-

Josh Robb:

Not point me to the answer.

Austin Wilson:

Yeah, instead of point me to the answer or show me a video about the answer. Just boom, here’s the five steps on how to reset your VCR timer.

Josh Robb:

That’s right.

Austin Wilson:

Because Josh probably still has VCR.

Josh Robb:

There’s five whole steps, man.

Austin Wilson:

You have a VCR?

Josh Robb:

That’s a lot. I don’t have a VCR.

Austin Wilson:

Okay. I don’t either. We have a DVD player.

Josh Robb:

I have a DVD player.

Austin Wilson:

We don’t use it ever.

Josh Robb:

Yes, it’s in the basement.

Austin Wilson:

It’s unplugged. Ours is actually unplugged. What’s a DVD? I watched some show and they said, “What’s a DVD?” So that’s funny.

Okay, so AI also could have an impact on education and learning. And this is-

Josh Robb:

Oh yeah, good and bad.

Austin Wilson:

Something we’ve talked about quite a bit. On the good side, it could have personalized learning experience based on individual needs and preferences can be generated through AI. So that’s a benefit. I have a special needs daughter, maybe this could benefit her. We could put in some of the ways that she learns best and what we wanted to learn. And boom, it’s going to spit out its personalized lesson plan. I don’t know. You could have intelligent tutoring. So you could have AI powdered-

Josh Robb:

You hope all tutoring is intelligent tutoring.

Austin Wilson:

No promises. You could have AI powered tutoring systems, which give students feedback and guidance automatically based on what they’re teaching.

Josh Robb:

Oh, that’s cool. Yeah.

Austin Wilson:

Yeah. You could automate administrative tasks through the school, grading, scheduling, record keeping. Which could really reduce the workload on educators. And then because this is a free open source kind of thing, it kind of democratizes a lot of what you want. Because you can say, I don’t have any resources, but if I have access to the internet, I can learn whatever I want, however I want, in real time. So there are also some really cool implications as it relates to space exploration. Because we can’t send a human to wherever. A, it might take our whole lives to get there, or B, we might just want to analyze it without actually going there. Well, AI can help with that. Number one, autonomous spacecraft navigation.

Josh Robb:

Oh yeah.

Austin Wilson:

If you don’t need to send a human to Mars, it really becomes a lot more realistic to get there. Because who has 10 years to be sitting in a space shop? Not me.

Josh Robb:

The one astronaut in the back, are we there yet?

Austin Wilson:

Are we there yet?

Josh Robb:

Stop it.

Austin Wilson:

So that is one way. Resource management, I mean, you need to optimize your resources when you’re looking at things like space, because energy consumption’s real. What do you do with your waste? What do you do with your food? Where you can use AI to kind of help you optimize that situation. You can analyze images through AI and that can really help you see certain geological features, landing sites, other important information, temperature, probably. Robotics, you can send robots and drones to different planets and enable more efficient exploration data, collection from there.

Predictive modeling, you could say, plan my space mission to Mars in this spacecraft. And it will be able to help you-

Josh Robb:

I’m ready to go help go. No, I got my backpack.

Austin Wilson:

Plan that out. Which will be-

Josh Robb:

I got my suitcase.

Austin Wilson:

More efficient, more cost-effective over time and really reduce the risk.

Josh Robb:

I want just a carryon, by the way.

Austin Wilson:

That’s good. You travel light.

So we’re going to hit on it one more time because this is kind of the key. As AI becomes more autonomous, more by itself, there is a risk that it may make decisions that are not what we would want them to make. That is interesting. AI could be vulnerable to cyber attacks, especially if it’s used in critical systems like healthcare, like transportation.

Josh Robb:

Oh, yeah.

Austin Wilson:

That could be a risk to public safety.

Josh Robb:

If you had artificial intelligence for stop lights because they’re just on a set time, they’re just on a timer. But if the stop light could learn traffic patterns and say, well, this is by a school. So from 7:00 to 9:00, when everybody’s coming and going, let’s adjust which direction is longer, they’re heading into the school versus the cross traffic.

Austin Wilson:

Because nothing’s worse than sitting at a stoplight when they’re no one there.

Josh Robb:

No one’s coming through. Yeah.

Austin Wilson:

AI could really help there.

Josh Robb:

But you would want somebody to take it over and mess with it.

Austin Wilson:

No, that could cause chaos.

Josh Robb:

That’s a security issue.

Austin Wilson:

And the final point that I want to talk about is just regulation. Because there really isn’t regulation over artificial intelligence at this point. But there is going to be some sort of need for regulation, that’s not where we are right now. And I think that’s a lot of the concern that people have with AI. So that’s going to be coming. Because AI’s advancing so fast. There’s a lot of people calling for a pause in development-

Josh Robb:

Like wait a minute.

Austin Wilson:

Like, hey, take six months. But the problem is that things like a place like China or Russia or North Korea or Iran-

Josh Robb:

They’re not waiting.

 

[32:10] – The Guys’ Personal Opinions on AI 

Austin Wilson:

They’re not waiting. So if we fall behind, we’re in real trouble. So I think that’s sort of the risk there. So let’s conclude by giving our opinions on AI. So I’ll start. I think it’s incredibly powerful. I’ve dabbled with it, I think it has a lot of really cool use cases. I don’t think it takes the place of a brain.

Josh Robb:

No.

Austin Wilson:

So I’m not worried about that. I do think if we rely too heavily on it, we’re going to go down the wrong path. I do somewhat have concerns about the legal side of it. This could really cause issues on that side of things. Who was responsible? I don’t know. The computer told me to do that. That could be a really big issue. Unlike some people, I don’t think we can afford to take our foot off the gas and development as our nation and our economy, because if we fall behind some of our less friendly people in the world, then they could use that against us. I’m not saying we want to use it against them, but we just don’t want to be vulnerable. And that’s the issue. I think there’s investment opportunities in the area, we’re not going to talk about that today. But a lot of stocks have been going bonkers on this topic already. So what about you, Josh? What are your thoughts?

Josh Robb:

When I first heard about this, and I’m thinking more on the ChatGPT side, is it seemed like a novelty to me. Like, oh, that’s cool little chat box you could type in and get an answer for. But you’re right in that the AI and this advancement learning concept can be very potential value or threat, depending on how it’s utilized. And I think there’s a lot of people trying to get in on it now without the understanding of what it could lead to. So yeah, I’m worried from that standpoint. But I think there’s enough people aware that there will be regulations to help at least eliminate some of that risk or worry.

Austin Wilson:

Absolutely.

Josh Robb:

I do think though, there are so many things that can be utilized here that’s to the vantage. I mean the healthcare one to me, especially, is if you can have that help for reviewing, looking through scans, and then someone interprets that and make sure it matches up. I mean, it just makes it so much better for that. Like that healthcare industry in particular. So I’m excited about that end of things.

Austin Wilson:

Absolutely.

Josh Robb:

Now, I also asked ChatGPT, what is your opinion on artificial intelligence? And this is what I got, you ready?

Austin Wilson:

Yeah.

Josh Robb:

As an AI language model, I don’t have personal opinions or beliefs as humans do. However, I can provide some information on AI. Then it followed through with some of the stuff you just talked about. But here’s the last little paragraph it wrote. It’s important for society to have a nuanced and informed discussion about the opportunities and challenges of AI, and to ensure that it is developed and used in a responsible, beneficial way.

Austin Wilson:

Hey, there we go.

Josh Robb:

Yeah, I mean, it’s like warning you itself. Like I could be crazy. Keep an eye on me. That’s what he is telling me right now. That’s what AI is telling me.

Austin Wilson:

And on that note, if you have someone in your life asking about ChatGPT or AI, send this episode to them because hopefully this will be able to help point them in the right direction. Again, we’d love it if you subscribe. As a reminder, if you maybe have questions about your financial situation, check out the Invest with Us tab on our website and you can get in touch with us and we can see if we may be able to help you. And leave us a review on Apple Podcast or Spotify, wherever you listen. We would appreciate that, as well. And as a reminder, we are pretty active on our social media, whether that be Instagram or Twitter or Facebook. So follow us on there and hopefully we can connect. So thanks for being here.

Josh Robb:

That’s right.

Austin Wilson:

ChatGPT, AI. Until next week-

Josh Robb:

Talk to you later.

Austin Wilson:

Have a good one. Bye.

Thank you for listening to The Invested Dads Podcast. This episode has ended, but your journey towards a better financial future doesn’t have to. Head over to TheInvestedDads.com to access all the links and resources mentioned in today’s show. If you enjoyed this episode and we had a positive impact on your life, leave us a review. Click subscribe and don’t miss the next episode.

Josh Robb and Austin Wilson work for Hixon Zuercher Capital Management. All opinions expressed by Josh, Austin or any podcast guest are solely their own opinions, and do not reflect the opinions of Hixon Zuercher Capital Management. This podcast is for informational purposes only and should not be relied upon for investment decisions. Clients of Hixon Zuercher Capital Management may maintain positions in the securities discussed in this podcast. There is no guarantee that the statements, opinions or forecasts provided herein will prove to be correct. Past performance may not be indicative of future results. Indices are not available for direct investment. Any investor who attempts to mimic the performance of an index would incur fees and expenses which would reduce returns. Securities investing involves risk, including the potential for loss of principle. There is no assurance that any investment plan or strategy will be successful.