Automation Bias: The Brain’s Trust in Technology

Automation Bias in Investing: When to Trust (and Challenge) AI | Bryan Foltice Behavioral Finance Podcast

In this episode of the Bryan Foltice Behavioral Finance Podcast, we explore automation bias—the tendency to use technology as a mental shortcut in place of vigilant information seeking and processing—using examples like GPS routes and copying ChatGPT answers without thinking. We discuss early research from the 1990s on pilots and automated aircraft, outlining two error types: omission errors (failing to act when automation misses a problem) and commission errors (following a bad automated recommendation despite contradictory information). We connect automation bias to Type 1 vs. Type 2 thinking, arguing that as tasks become more complex and consequences rise—especially in investing, retirement, tax, and estate planning—we need a middle ground: use AI’s benefits while actively verifying, prompting, and challenging outputs. We close by previewing next episode’s topic: algorithm aversion.

00:00 Podcast Welcome
00:26 Automation Bias Defined
02:19 Origins in Aviation
03:47 Omission vs Commission
05:39 Why We Love Automation
08:04 GPS Jog Gone Wrong
11:35 Type Two Thinking Trigger
13:07 Finding the AI Middle Ground
16:08 Prompting and Staying Critical
17:57 Algorithm Aversion Teaser
19:37 Wrap Up and Next Episode

Transcript

Copy Transcript Copied

SPEAKER_01 0:04

Welcome to the Brian Fultis Behavioral Finance Podcast, where we unravel the mysteries of behavioral finance and unlock the secrets to making smarter, more informed decisions with your money. Now, here's your host, Dr. Brian Fultis.

SPEAKER_00 0:24

Welcome back, everybody. Have you used your GPS on your phone and it has taken you into road construction? Or have you used Chat GPT and copied and pasted its answer without giving it much thought? If you've done those, and I have done both, you have fallen to what we're going to learn about today called the automation bias. And so the automation bias today is this tendency for us to use automation. So in our world, we have technology right in our smartphones to use this automation as a heuristic or a mental shortcut replacement for vigilant information seeking and processing. Okay, so that's our definition of the automation bias, the tendency to use automation as a heuristic replacement for vigilant information seeking and processing. Now remember, when we talk about heuristics, these are called, these are our mental shortcuts, things we do to preserve mental energy that usually have a tendency to take us in a shortcut direction in getting us a sufficient answer without really overthinking it or overburdening our mind. And this helps us conserve energy. So we're going to talk a little bit more about that. But just so we know, that kind of mental heuristic is also known as a mental shortcut. And we see that with our technology. So essentially, we're using technology as a mental shortcut to actually do our thinking. And we're going to talk about when that works to our advantage and then also when it doesn't. And when it doesn't, what do we do about it? What can we do to improve that? Okay, so before we jump into all of that, we're going to start at the beginning here because this is a very relevant 2026 topic that we're talking about. But this automation bias didn't spring up recently. It's been around for a while. And when I look here, I'm using the 1999 Proceedings of the Human Factors and Ergonomic Society's. I don't even know what that means. 43rd annual meeting in 1999. And the recap that I have here in front of me is from Kathleen Moser and Linda Skitka. And from what I can gather from this summary is that they were on the front lines of this automated automation bias. And others had contributed to this in the 90s, but this looks like enough that this was probably their research program during their time. I don't know if they were PhD students at the time or if they were young assistant professors trying to get tenure. Nevertheless, they wrote a really nice summary on this automation bias. And what they were doing back in the 90s when we really were formalizing this is they were using pilots using the technology in the aircraft. And so they had automated aircraft and they were talking about pilots, and they broke this automation bias into two different types of errors. And one type of error is called an omission error. Okay, so the omission error is defined as failures to respond to system irregularities or events when automated devices fail to detect or indicate them. In other words, the computer does not tell me there is a problem. I go along with it because I'm waiting for the computer to tell me that there's a problem. This is our omission type of error. And I didn't do anything because the computer never told me. Omission error. Still using automation bias in that way of omitting the error. Type number two, automation bias error, is the commission error. This is when, as they define it here, when people incorrectly follow an automated directive or recommendation without verifying it against other information or in spite of contradictions from other sources of information. And this is basically when the computer gives you the error and you co-operate and go along with the error that they've given you, and you never question it. You just go along with what it said. That is co-mission errors. Okay, so now if we have both types of errors, omissioning and co-mission, we want to understand why we fall into this, why both types of errors play into our biases. Well, our brains love automation. We love a nice break in any time we can pass off our brain workload, and in this case it's so easy to pass it off to technology, we're gonna do so. And and so this is what we talk about when we talk about type one thinking. Type one thinking, this is our fast and slow types of thinking you might have heard about. Type one thinking is our automatic thought process that we can go do our morning routine without giving it much deliberate thought. I can do my 35-minute commute down to campus without giving it much thought. Almost scary how much, how little thought I give. And you'll make it down the road and you go, oh my god, what happened? How did I get here? I don't even remember the last 20 minutes. Automatic. Now, the good news is, thank goodness, it keeps you alive. Your your type one brain keeps you alive, but it also conserves mental energy where you're not really overthinking this. Now, what we do with our smartphones here, we fast forward is we let them do a lot of our thinking now to the point where I know how to get to campus from my house, but I still check now, just in case maybe there was an accident, and that I lean into the technology to help me find the most direct way, and just pass that off. Something happens, the phone will tell me. And so this is when it can potentially become dangerous. Usually it works, but at times it's not a hundred percent accurate or true, where it might tell you where to go where everybody else is going. So now all of a sudden you think you're moving away from the accident, but now you're going on this alternative route that the same computer has told everybody else to go, and now you're just in the another traffic jam on the side of the road, on the trying to get around anyway. You you can see it, it doesn't always work. And when I was just got back from Charlotte, North Carolina, we were at the higher education financial wellness association called Heftwa. And I was on a panel there in Charlotte, and on Sunday I went and worked out at the fitness center that I have a membership at, which is kind of this branch regional thing, and decided that after my workout, free day, I'm gonna instead of taking Uber there, I'm gonna jog home. And it said, four miles, you can jog home. Here's the route. I said, okay, make sure you don't take me on this route that would bring me by this really busy interstate. Obviously I'm running, so just leave me out of that. And it said, okay. So I am jogging my way and having a really nice jog, and it's weaving me through all the back streets that I had asked it to do, and then suddenly there was a take a right at the path. Okay, no worries, path. This path was overgrown with different types of vegetation, and there was this real small path that I was trying to get down. So now, put yourself put myself in your shoes. What would you do with that? Do you take the path long? GPS tells me this is where we need to go in order to get back to the hotel in the shortest way possible. And I'm going, you know what? I gotta do this. I don't know any other way. And so I'm gonna take this path, and I hope that I can come out through this back end without having my body cut into 30 different pieces at the end of this path or somewhere in between. And it took me through this really narrow path, and then there was a bridge, and we could cross the bridge, and it was telling me where to go, and it was a little muddy, and I'm like, oh uh now I'm committed to this, I'm totally in. And and then we get to the end of it, and it shoots me out into the street. We say, Oh, okay, almost there, just got to get to this final road, make a right, and I've got a mile to go. And by now, it it's North Carolina heat in the summer, and I'm really sweating, and I'm ready to get back to the hotel. And sure enough, I take a left, 0.9 miles to go, and it has me right on that interstate. And so now I'm stuck. Now I have 0.9 miles to go down this interstate, and I'm trying to just not get arrested, and at the same time, not get killed by these cars that aren't expecting a guy jogging. And so I just was super motivated to get through there, but again, I finally made it, and I was like, the damn GPS. I told it not to take me down this road. I was totally, I chose it a little bit further out to do this, and sure enough, it kicked me back, and I had no other choice but to take that type of route. Anyway, when is it time to move into type two thinking? So, again, when will these shortcuts hurt us? And when I teach behavioral finance, it's all about understanding our shortcuts, our mental shortcuts, and then also understanding when is it time to tap that adult thinker, that critical thinker, bring the adult into the room to have these conversations. And so if we're just doing day-to-day little tasks and we want the ChatGPT to help us create an email, then okay, that's fine. But as the consequences get a little bit more significant and the task gets a little bit more complicated, this is when we need to really come back and think about is it time to hire out our type two brain and let us really start to think through this. Now, the problem is most people think in binary terms that I'll never use Chat GPT to help me, or I'll never use technology to help me through processes. And on the other hand, you have people who will say, I will use them for everything, and I'll just follow this blindly. And what we always try to tie in in our conversations is it's never black and white. It's never that easy. And so that's what we're trying to mesh together here is when is it convenient to use your type one thinking? And then when is it time to move into type two thinking? And so obviously, when we're talking about our finances, when we're talking about our investing, our retirement planning, tax planning, state planning, all these different things, yes, we we have a choice, right? We have these two goalposts, never use Chat GPT, never use AI, or use it blindly. And what I'm trying to say is let's see if we can find a middle ground, find a place where you go, okay, this is good, but now I'm gonna, I need to challenge these things, these particular aspects here, or I need to confirm these aspects here on what is the right way, what is the actual best way to go. So again, we're using the best of both worlds, still needing to use our critical thinking, needing to use our brains, um, but also using the the benefits of this technology that we have today. And I find myself falling victim to this, is why I study this, why I find it super interesting, because I've noticed myself super easy tasks, like a literal one or two sentence email. I'm asking Chad GPT to help me write that. And I've had to just laugh at myself, going, why? Why just write it? Use your brain. You can still use your brain. Um, but then you realize how easy it is to just have Chad GPT help you out, or Claude, or whatever your your AI learning system is. And and so we have this right in our phones, our ability to use this, have this automation bias and just have the computer tell us what to do, and we become the computers without really thinking about what what we're really doing. And and so anyway, here we're here's what we want to do. Our takeaway is automation bias is not about computers. Um, it's actually about us. So we've we're basically just showing another shortcut, a mental shortcut that our brains take. Just happens to be technology that is providing that opportunity for our brains to just kind of sit back and take a vacation. And and so the question is, should I use AI? That that's not the case. It is how do we have this balance of how much should I trust AI? How much do I need to think on my own? And this is where I'm just it, I don't know that middle ground. I just know that it's not one or the other. But when we start thinking critically or keeping that critical thoughts, we can use that power of AI to help us really enhance our lives, enhance our thinking, and and make our lives better. And this is where we can even talk about prompts. How well are we prompting AI to give us what we need? So this again becomes something that is on us, something that we have to critically think through, but then we can pass that over to AI. At the same time, we also have to remember that we have to reinforce this. I've noticed, I've told it to be more critical and to ask me questions back. And AI has been really good at at challenging me and having more back and forth, but without the reminder every couple of days, it's like having like my teenage boys, it's having to remind them. Hey, remember, we said we're gonna be critical. Oh, yeah, yeah, that's right, that's right. We still have to have that kind of back and forth when even dealing with AI, so I think of them almost as same as teenage boys. Once you tell it once, it's not going to remember right away. But hopefully over time it can still help us think better without forsaking our ability to think critically. Now here's the closing thought. Best investors won't be the ones who ignore AI. Um, they won't be the ones who blindly trust AI. We've said that very clearly here in this episode. They'll be the ones who know exactly when to challenge it. That's what I want to think about. I want you to think about. When is it appropriate to challenge? When do I need to step in? When do I need to re prompt or reprompt my commands to give this something that is actually going to help me, not just give me a mental crutch, which is what these mental shortcuts can be. But this is where it gets really interesting. So if people trust automation too much, then you'd expect the obvious solution is trust it less. Except that's not what we found in research. In fact, many people, if they don't trust it, they swing all the way to the opposite, just like we talked about before. Now, the moment the algorithm, and we ask them why did you swing from one to the to the other, and we found out in research, um, I didn't do shit in the research in this area, I'm just the the person who's connecting this. They found that when AI or the algorithm made a mistake, they had a really strong adverse feeling about it, and they went the complete opposite direction. So when that algorithm made the mistake, they went in the way opposite direction. Even though the algorithm is still statistically more accurate than they are. So when we come back to this next episode, we're gonna try to dissect here, okay, what is the bigger behavioral mistake? Trusting the computer too much, like we talked about today, or not trusting it at all or enough. And that's what we're gonna pick up next time as we explore one of the most fascinating findings here in behavioral science. And this is gonna be called the algorithm aversion. Okay, so we're gonna leave it at that. And I want you to come back next week if we tie up this conversation on the algorithm aversion. Hope you enjoyed this episode. If you have any questions, please just click in the response right below that will take you right to either a voicemail if you want to do a voice message or a text. If you have any comments or ideas or questions, starting to get some questions, and it is really exciting because we can start doing some QA podcasts and really start a legitimate dialogue with this really interesting conversation. Check out the show notes for the link to our website, also the newsletter that we have out now where you can stay up to date on everything that we're talking about, but see it also in newsletter form so you don't really miss anything that we've been talking about here. So we're just getting started on our reboot here. I'm super excited to continue this. I can see this vision of really seeing how we can make this popular, we can make this interesting and engaging, but I also want you to feel the same way and engage with this as well. Love to hear from you. In the meantime, hope you have an awesome rest of your day. We'll see you in the next episode. Thanks, everybody.

unknown 20:56

Bye.

SPEAKER_01 20:59

Thank you for tuning in to another episode of the Brian Fultis Behavioral Finance Podcast. We hope you found our exploration into the fascinating world of human behavior and finance, both enlightening and thought-provoking. Be sure to subscribe for future episodes. And until next time, stay curious and financially savvy.