Is AI Your Partner or Competitor?

How is artificial intelligence changing the field of accounting, and should we be concerned that AI will replace entry-level positions in this field?

We explore those issues and more in this episode of Inside Alvarez Business.

Abby Parker is the Nancy and Frank Kudla Fellow in Artificial Intelligence and assistant professor of accounting in the Alvarez College of Business. Her work explores the intersection of AI and accounting, and her recent research utilizes AI to predict accounting errors. Abby sees AI as a tool which can be used by accountants to improve the field of accounting, facilitating the tasks accountants engage in without necessarily replacing the accountants themselves.

Hear her views on how AI will impact the field of accounting.

Stay connected with the UT San Antonio Carlos Alvarez College of Business to learn more about how we are empowering the next generation of business thinkers. Follow us on social media or visit us online at business.utsa.edu

Transcript

Chanyuan (Abigail) Zhang Parker, PhD 0:00

I think in the world of AI, we need to highlight the importance of accounting because accounting is a profession of trust, right? So in this world of AI where we worry about hallucination from AI or misinformation, we as accountants actually have a bigger role to play and show that we can demonstrate accountability.

Jonathon Halbesleben, PhD 0:26

How is artificial intelligence changing the field of accounting? And should we be concerned that AI will replace entry-level positions in the field? We explore these issues and more in this episode of Inside Alvarez Business, a podcast dedicated to bringing you the stories behind the research. I'm your host, Jonath on Halbesleben, Dean of the Carlos Alvarez College of Business at The University of Texas at San Antonio. Our guest today is Abby Parker, Nancy and Frank Kudla Fellow in Artificial Intelligence, an assistant professor of accounting in the Alvarez College of Business. Her work explores the intersection of AI and accounting. In her recent research, she utilizes AI to predict accounting errors. Abby sees AI as a tool which can be used to improve the field, facilitating the tasks that accountants engage in without necessarily replacing the accountants themselves. I hope you enjoy the conversation as much as I did. She joined the college in 2023 after completing her PhD in accounting from Rutgers. She has the unique distinction of being the first faculty member I hired when I became Dean. So I will always remember you for that Abby. She's just getting started with her career, but she's already published groundbreaking work at the intersection of artificial intelligence and accounting. Abby, welcome to Inside Alvarez Business. I'm really looking forward to hearing the stories behind your research.

Parker 1:42

Yeah, thank you so much, Dean Jonathon. I'm so glad to be here.

Halbesleben 1:46

So just to get started, we'll talk plenty about your research in a minute, but just to kind of get started a little bit, I'm always really interested in how people decide to take - if you go into an academic career. It's not very common when people are growing up to say, I want to be an accounting professor. So how did you get into this field?

Parker 2:05

Right. So I think when I was in high school, I actually wanted to do biology, like things that are more science-based. But then my dad talked to me and he said, "Have you considered like getting a job, like an actual job after graduation?" [laughter] I said, "Yeah, that's a good point." And then he said, "You know, like being an accountant is a great profession." And so then I was like, yeah, maybe. So then I just decided to change my path, and then I did, I actually majored in finance and minored in accounting in my undergrad. And in my undergrad, I just feel I probably am on the nerdy side, and I did a couple of internships at the banks and I didn't like it. And I feel maybe I just want to continue studying. So that made me pursue a master's degree.

Halbesleben 3:02

Okay.

Parker 3:02

And then during my master's degree, I get to know a faculty who is doing research related to accounting information system. And at that time, he introduced me to this company that is digitizing their accounting software. So, and then from that project, I feel the digitization technology applied in accounting is going to be the future. And so then I decided to make this my career path. And then I had this opportunity to do a PhD in accounting information system at Rutgers, which has the number one ranking in this field in the United States. So I just say, "Okay, I'm leaving, I'm going to do a PhD." So here I am, and I love being a faculty, being in academia. So I what I love about academia is that I get to set what I am going to do myself. I hate being told to do something. There's something about me. I just want to have my own initiative, and the process of putting together a research paper is fabulous because it's like building an art from ground up. So when the moment you have an idea and then you go to collect data and then see whether there's results, and you fine-tune, and then you draw in the theory. So the whole process can be long, but it's very rewarding. And to me, it is such a joyful process. So I love being in academia. And also, teaching is also another very, very important aspect because, like, whenever I walk out of the classroom, I don't feel tired, I feel recharged because the way you communicate with students and then you receive their feedback really feel make me feel that I'm doing something important, that I am giving them something that they can use. So I just love being a faculty.

Halbesleben 5:02

That's great. Well, judging by the energy you bring to this conversation, I can only imagine how much energy you have in the classroom [laughter]. So I'm excited for you. So we're gonna talk some about some of your work in artificial intelligence and accounting, but you know, you've really been exploring some of this stuff for for quite a while now, really before that was even talked about much in the media. So you've been studying issues like robotic process automation and machine learning, like prior to the pandemic, even. So before we get into the specifics of your research, how did this area become something that you were interested in looking at? I mean, you kind of hinted at a little bit with your career path. But I'm curious, like, how did you settle on this stuff?

Parker 5:41

Yeah, so yeah, as I mentioned, like during my, I think it was 2016-ish, when I was in my master's degree, I learned about digitization and I know that this is going to be the future. So then I came to Rutgers. So let's call it FATE, because when I joined, at that time, this term robotic process automation was the popular term. It's similar to the AI that we use now. And then at that time, my advisor just said, "Hey, Abby, we have this collaboration with a company, this accounting firm. They want to figure out how to automate their routine, like a mundane tasks using RPA. Can you please take a look?" And I said, "Sure," like you can throw me anything because I'm here to learn, right? So then I got into this field, I started digging, and then I got very interested in this field. So then that's how I stumbled into this field.

Halbesleben 6:34

Okay.

Parker 6:35

And then I happened to love it. It's funny when you say that you say, oh, you started early, but at that time I feel that I was already behind because the technology world is really changing every day. So whenever I wake up, there's a new term, there's a new application. So I was just trying so hard to catch up. But unfortunately, I think at that time in the whole business world, people are not so familiar with the technology. So it's interesting that actually in 2019, I had my first paper on AI, and we sent it to accounting journals, the top journals, and then we got the feedback saying, what is artificial intelligence and why should we care? We don't buy the story. And it got kicked back. And two years later, or three, four years later, Chat GPT came out, and then everyone says, "Oh, your data set is outdated because you are not using generative AI, and generative AI is the only AI." So, yeah, it's fun to see how the perception has changed so much, but actually the work we have been doing or I have been doing is always there, and I hope that the general audience or readers can appreciate that. Yes, AI is, you know, new, but it's also not that new. It's been there for a while. It's just that the recent development has brought that to our attention. Yes.

Halbesleben 8:00

It it's so interesting. You know, I hear this from from researchers in a variety of different areas, how challenging it can be to sort of keep up, you know, since research sort of by definition looks backwards somewhat, trying to keep up with the technological changes, so cybersecurity and some of these other areas, it's similar.

Parker 8:18

Right.

Halbesleben 8:18

So I can only imagine how challenging it is for you to kind of keep up with all of that. And yeah, especially with the the length of time it takes to conduct a good study and get all the data, and then of course go through the whole publication process. Must be kind of frustrating at times.

Parker 8:34

Yeah, so I just I know that we're deviating a little bit, but this is part of the reason why for some of my publications I need to be fast because I am looking at the intersection between technology and accounting. And the traditional accounting research is slow by definition because they are looking at the very fundamental or questions that may not be so time relevant. However, the technology issue is like, okay, now if you publish a paper on RPA, it's outdated.

Halbesleben 9:04

Yeah.

Parker 9:04

So then I have to be aware of this timeliness. So I need to be very strategic about okay, which paper I am willing to sit and wait for, let's say, something to play out. Or, other papers are more timely, and I need to just get it out and get more citation. So there is a balance. So that's why if you see my profile, I have something that's really quick, but on a very timely issue. But I have something that's more, it's like a wine that is [laughter] that is doesn't cost a lot to sit there for a while and let it let it simmer. So yeah, so that's something about the type of research that I do.

Halbesleben 9:45

Okay. So you mentioned some of your really early work on robotic process automation. Can you kind of describe what RPA is and like how that differs from some of the more current like generative AI models?

Parker 9:57

Yeah, so robotic process automation is basically a piece of software or application that you can install on your computer that can help you automate some really rule-based repetitive tasks. For example, if you do certain things every day, like you open this folder and then you move this file to another file, and as an accountant, they do this kind of work very often, right? So this kind of software is able to automate that. But then, okay, this doesn't sound very interesting. But the new thing to RPA is that it the interface has made, has been made so easily that for any people who don't know about technology, they can just drag and drop, play around. It's a little bit like Chat GPT, how easy to use. But like they can also use it to build their own automation. So this is feature number one. Feature number two is that this kind of software can mimic human action. Like if you click this button, then the software is able to mimic you clicking that button. So it's not like you using a Python programming or a SAS code or a stata code to do hard programming. It's basically mimicking what you do day -to -day. And this is especially charming to accountants because accountants work with these accounting software that doesn't necessarily have Python code to automate, right? You have to go through the interface where you log in using your username and interface, you click this button, you copy paste numbers to the software. So RPA just very nicely walk into this field and to do what we call last mile automation.

Halbesleben 11:30

Oh, okay.

Parker 11:30

Yeah. But now what's the difference between RPA and agentic AI? So RPA, let's call it dumb robot, because RPA can only automate things that you explicitly tell it to do. Like first open this browser, then type in this search bar and what exactly to type, and then click the search button and so on and so forth. And not only do you need to build out all the steps, you also need to tell it how to react in exceptions. Like, for example, oh, if this browser takes too long to respond, should I quit? Should I do something else? So RPA is very rigid, it only takes on rule-based tasks. Whereas a gentic AI, you can think of it as a large language model powered RPA, where it can just, you don't need to tell it what to do, you just tell it, hey, make me a restaurant reservation. And then in the background, the generative AI is able to automatically generate the steps that it needs to do and directly handle exceptions. So one is passive and rule-based, another one is more autonomous and take action on its own. So that's the difference. So the latter is just more intelligent, the former is dumb or rudimentary.

Halbesleben 12:52

Yeah, yeah. It's interesting when you describe it that way, you can kind of see the progression from like hard-coding things and really basic if-then statements.

Parker 13:00

Yeah, yeah.

Halbesleben 13:00

You know, up now into the more agentic models that we have now. So that's really interesting.

Parker 13:05

Right.

Halbesleben 13:06

Great. So let's talk a little bit about you've had a paper come out not that long ago now in the accounting review, one of the top journals in the accounting field. And in that paper, you used machine learning to model, I'm sorry, to predict material misstatements. To start, can you take a little bit of time, tell us what a material misstatement is so that we can understand the problem?

Parker 13:26

Yeah, yeah, yeah. So for people who don't know too much about finance or accounting, so the public companies they issue their financial reports, which we call 10-K reports. So generally people just assume that, okay, this report is reliable. If a company says, let's just make it up, hey, we have revenue of $100 in the past year, you just believe it, right? But the reality is that the process of putting together a financial report is complicated, and how companies are arrived to that number has many, many steps and judgments to it. So the managers who you know provide the report, they will make mistakes, either intentional or unintentional. And then you have auditors come in to kind of check if the financial report has error, but auditors sometimes cannot really capture the error. Again, either intentionally or unintentionally, right? So you end up with this situation where the financial reports that you get from the SEC, the 10-Ks , may contain errors. Now, for these errors, the error might be small or might be big, let's say for this $100 revenue case. If the company later says, "Oh, sorry, investors, actually our revenue should be $99." And then as an investor, you'll be saying, "Okay, it's only $1 difference, right? 1%. It doesn't matter." It doesn't change how we perceive the, you know, your growth, your fundamental. So it's not a big deal. But if the company came out and said, "Hey guys, sorry, actually we only made $50, right?" So then that fundamentally changes how you perceive the profitability of the firm. So this kind of mistake is called a material misstate. So just to summarize, right, material misstatement is just something that is material enough that would change how investors would perceive or value your original financial report. If that fundamentally changes that perception, then we call this a material misstatement. So.

Halbesleben 15:23

Those are publicly filed.

Parker 15:24

Yes.

Halbesleben 15:25

Yeah, so that people of course...

Parker 15:27

So people can know that, yeah. So when this happens, the companies will issue a formal statement in Form 8-K to the SEC saying that, "Hey, we realize that we have this mistake. And please investors do not rely on that a faulty financial report, and we're going to reissue you a modified version." So that's how the public knows that which financial report is faulty and not to rely on them.

Halbesleben 15:52

Yeah. So it seems like if I'm, if I was reading the study correctly, and please let's not assume that I read the study like correctly, that the main innovation here was it previous work had been able to sort of go backwards and explain why a material misstatement happened. Your model allows it you to sort of predict in the future, based on these conditions, we would expect a material misstatement. Is that the practical thing?

Parker 16:20

Yeah, exactly. That's exactly the meat of the paper. So as you pointed out, the prior work focused on detection. So what is detection? Let's say Apple issued their 10-K in 2025, right? And now for this 2025 financial report, we look at whether it has a mistake. Now, this is called detection because you are identifying errors in a financial report already issued. And I call this reactive because by the time you identify error, sorry, the damage is already done, right? Because the false information is already in the market. And then now investors need to kind of pull back their investment and there are other consequences. Whereas my work focuses on forward-looking prediction. That is, okay, if Apple is going to issue their 2026 financial report, how likely is that report going to have a mistake? So we are moving from reactive to proactive. And this is especially important for investors because investors, especially those that focus on their long time, long-term strategy, they want to identify some opportunities that the market has not exploited, right? So then we can kind of proactively see where the risk is, and then either you take advantage of the risk or you just avoid the risk. And so, yeah, that's the core part of the paper. We're just going like one step ahead, do a prediction.

Halbesleben 17:54

Were there like consistent conditions or kind of features of companies that you see consistently would predict these kind of misstatements?

Parker 18:03

Yeah, you mean like what are the features that...

Halbesleben 18:06

Yeah.

Parker 18:06

So we did some feature importance analysis using some like a very fancy term called explainable AI because we build machine learning models to make this prediction, and the machine learning model is considered a black box. So we just try to use some tech method to like open the black box and see what features are important. So I think we what we found is consistent with what prior literature found with some extension. For example, we found that audit-related feature, for example, the auditor tenure, meaning how long the same auditor has been with the company. So and we find that this relationship is not linear, and this is very consistent with what prior literature has hinted. So, for example, if I audit Brittney, all right, like for the first year when I audit you, I'm not so familiar with you. So I may miss some important things, and that increased the chance of having a misstatement. But then when I'm with you for like a few years, we are pretty comfortable with each other and I'm familiar with your situation, and then that decreased the risk. However, if you increase again, let's say I'm with you for 10 years, then that caused some issue because we may be too familiar that I am too afraid to speak up to you that, hey, you need to correct this issue, right? So when then we see increase of risk again. So that's one important thing. And we also found some new factors, but it's also very intuitive, which is this abnormal tax penalty. So there's not much research to it, but my interpretation is that that penalty only comes in when the regulators found some attack, like violations of the tax position of the firm. And usually the this kind of penalty is related to the complexity of the firm, and the complexity of the firm is just naturally related to the propensity of having more mistakes. And , we just caution the reader that, hey, these are associations, not causations. Even if you can find some predictive factors, you cannot say that, hey, because of this you have mistakes. We can only say that these are highly associated, these are just predictors. So, yeah, we do find some interesting predictors.

Halbesleben 20:21

Okay, cool. And this is the I think the first time we've had a podcast where someone's referred to the the producer, Brittney. So for the listeners, Brittney's sitting here in the room with us. Yeah. So, so taking this a step further, and I guess you kind of answered this a moment ago, but in some ways, and of course, these these misstatements aren't necessarily like fraud or you know intentional, but but they could be. And do you have any concerns about kind of predicting like what some could argue is bad behavior and making sure that's accurate? Or I'm thinking back to that movie, and you both are are younger than I am, but that movie Minority Report, where there was the the sentient beings who could predict a crime and then they would intervene so that the crime didn't happen.

Parker 21:09

Right, right, yeah.

Halbesleben 21:10

You have any concern about that, or is it

Parker 21:13

I'm not too concerned about that because I think the concern is well taken though, but I think the difference is that we worry about this kind of minority report or this issue when it comes to human beings, right? And then we talk about like ethics or bias and so on and so forth. But here we are dealing with public companies, and their job is to make sure that they issue, let's say, fairly represented financial reports that reflect their fundamental activities, right? So we like if we are able to build a model, right? And by the way, anyone with appropriate skills or the data set can build it, we can actually hold them accountable. And also it can be used for multiple stakeholders. I mean, even for if we take management as like a whole, then we can even include like the let's say audit committee, like just lump them into this big management team. So they also have this incentive to prevent fraud. So I'm not too concerned. And by the way, I think in the business world, we have tons of such scoring. For example, bankruptcy. We also have models to predict bankruptcy or other bad behaviors, quote unquote behaviors, right, of the firm. So I don't think it's we should equal put an equal sign between the like what we predict for corporate misbehavior versus human misbehavior. And I think for corporate, especially for public firms, we need to hold them accountable to some extent.

Halbesleben 22:45

I think that's a really thoughtful way of thinking about that. I mean

Parker 22:50

Yeah, and also the investors, they yeah, the in the finance world, people are already doing that indirectly or indirectly because the investment, banking, the they are already developing machine learning tools to just directly predict the return, right? And to some extent you can also say, hey, predicting return high or low is also like a minority report because if you have a low return or low performance, is that a, you know, just the predicting some bad consequences? No, because it's all fair game. It's all public data. You can just exploit it and use it.

Halbesleben 23:23

Yeah. Yeah, you'll have to apologize or forgive my, of course, as a psychologist, my natural bias is because things are all tied to some human behavior.

Parker 23:31

That's also true, yeah.

Halbesleben 23:32

Yes, I do think it's interesting that kind of the way you way you've put that. I agree with you. It is it's not equivalent to think about it like that.

Parker 23:40

As a human, yes, yes.

Halbesleben 23:42

So kind of going in a little bit different direction. So a lot has been made of how AI is going to replace jobs, especially entry-level positions in accounting. That's kind of one area where you see this a lot. You've got a paper coming out in the Journal of Accounting and Public Policy that seems to make a really important point in this space. You found that machine learning models they may make fewer mistakes than a human auditor, but the models, like machine learning models and humans, they make different kinds of mistakes. And so really the takeaway is that they should be working in tandem. Can you talk about what that looks like in practice?

Parker 24:16

Yeah, yeah. I think actually our reviewer asked the same question. And so we had to add a discussion section there. So, yeah, so in our paper, because we cannot directly speak to how it works in reality, we can only say that, hey, if we combine the data that comes from machine and the data that comes from the auditor, and let's mix it and see what this data is about, and we just say, hey, it is just better than either machine learning alone or auditor alone. But how they are mixed together, we cannot speak to that. And but I can offer some of my own personal insight. So I think in reality, if we talk about the type of work where, you know, it's really judgmental, let's say risk assessment. So I'm just for audience that are not familiar with the accounting work, especially the auditor's work. So let's say I'm again, Brittney, I'm sorry about picking you again. Let's say I'm auditing Brittney, and whenever I start my work, I need to look at her, let's say, track record of revenue, whatever, and then I'll say, oh, how what's the risk of Brittney materially misstating her financial report this year? So in this kind of risk assessment, I need to, I as an auditor, I need to apply some of my professional judgment, either from my past experience in the industry or just auditing Brittney throughout the years. My preferred strategy is that the auditor I come up with my own risk assessment on my own without any assistance of AI to begin with. And then I can have AI to come in, utilize maybe the same set of information, or let's say some information that is too large to be processed, let's say a lot of like contracts, documents, right? AI digest, and the AI comes with its analysis together with some points that I need to be thinking about. And then I step back and look at both my original assessment and AI assessment and then see, okay, does AI make a point here? Does it provide me with some additional points I should consider? Or, you know, I don't I don't trust AI's judgment. I think my assessment is better. So I have to make the final call, the final judgment here. So that's my recommended approach for at least several reasons. The first reason is that I think - so when you ask AI to directly give you something, we, I think the psychologist here, you may know more theory than I do. You the human mind will be biased or constrained by this limited set or a cascading effect. So, and especially if AI is misleading, you cannot jump out of the loop. So I want the human to step in first and then chime in the AI and then do some more work. And yeah, I guess I said several reasons, actually just one reason. So yeah.

Halbesleben 27:12

No, it makes sense though, because I mean I could see how if you start with the AI, then you get anchored to what it has said. So you could get into some of the old anchoring and adjustment biases that we have. So I think that does make a lot of sense to have the human work first and then use AI to maybe verify or enhance the human.

Parker 27:31

Yeah, yeah, yeah, yeah, exactly. And , and that's also how I think I know we're going to talk about this later, but I think for us as human beings, we need to kind of hold on to this value that our brain can provide. Because if we just rely on AI to give the answer gradually, we just lose the ability to think for ourselves. So I think it's just important to kind of just treat AI as a tool. It's not it's not going to do all the important work.

Halbesleben 28:02

Yeah. So I mean, in some ways you've kind of previewed what I was gonna ask you next. You know, as we think about AI possibly replacing jobs, it's a tricky time for students because they can't necessarily plan if they don't really know how those are gonna how AI is gonna impact their job. So realistically, how I mean, and I know we only have so much information right now, but I mean realistically, how much do you think it's gonna impact the accounting job market in the next two or three years?

Parker 28:31

Yeah, so I know you sent us this article from Wall Street Journal where it says, I think it may be PwC or one of the big four accounting firms says 20-30% of their entry-level jobs or the just tasks will be automated by agentic AI. I believe that. I take that. But beyond that, I don't think AI is capable or should be able to replace the remaining work. So I classify accounting work, or actually this applies to all white-collar tasks into three types. One is called the semi-structured and then semi-structured, sorry, structured, semi-structured, and unstructured. So the kind of tasks that can be replaced by AI, either the rudimentary RPA or agentic AI, they belong to the structured tasks, right? For example, I as an accountant, I need to verify documents. We call it a three-way matching, meaning like from the invoice and the payment and also another document. So, like this kind of three-way match is very rule-based, mundane, nobody wants to do it. So it's great. AI takes it, no problem. Now, for the remaining task, why I say that I don't think AI will completely replace, at most, AI can augment, is that accounting is actually very complex. The situation varies so much depending on the client situation. It's almost like every patient is different. You have the, let's say we have the same cold symptom. Your doctor looks at you versus your doctor looks at me. They will say, I'm sorry, you guys have different physiology, different demographic. So the prescription is different. So the same applies to accounting and the law. So, like in those situations, AI is not able to have that kind of, let's say, or yet, right? That ability to adapt to very client-specific situation. And that's where human expertise is really important. So, and I speak to many professionals and they just agree with me on this point. And actually, I think in the world of AI, we need to highlight the importance of accounting, but not too narrowly as the accounting work that we do, but what accounting represents. Because accounting is a profession of trust, right? So if you think about it, we measure the business and we provide assurance on that measurement, and the public trust the numbers that accountants publish. So, in this world of AI where we worry about hallucination from AI or misinformation, we really need to play a bigger role to step in and say, hey, we as accountants, we actually have a bigger role to play in this whole in this era of AI and show that we can demonstrate accountability. So I think we need to advertise accounting in this more broader sense.

Halbesleben 31:43

Yeah. I like too that you've kind of made this distinction and I'm starting to see this a little bit more between the jobs and the tasks.

Parker 31:53

Yes, exactly.

Halbesleben 31:53

In some ways, I think it would help all of us to be more careful in how we talk about what AI is replacing.

Parker 31:59

Yeah.

Halbesleben 31:59

Because it - if when, for example, if it's 20-30% of the tasks, that's a dramatically different interpretation than 20-30% of the jobs. And so I'm actually glad you're kind of highlighting that. I think that's something that in general, not just like business schools, but just kind of in general in society, we need to be a little bit more careful in how we're talking about.

Parker 32:21

Yeah, exactly. And in this world, just the the professionals, they will find other things to do to compensate for that loss 20-30% of tasks. Like, for example, when I talk to one of the firms, they are using AI to kind of automate some of the work that couldn't be done because they cannot hire enough people. There is an accountant shortage, right? So now they can use AI to do it, but it doesn't mean that they don't need to hire any more people. Now they have more work to do because now they realize that oh, there's another business opportunity here, we can expand on this. So then they still need people, they just don't need people to do those kind of tasks.

Halbesleben 33:06

Yeah. So yeah, I do find it kind of interesting. Accounting in particular is sort of a unique field in that for the last, you know, five or so years, we've had employers begging us to produce more accountants because they needed accountants. And now suddenly, all of a sudden, oh, we're just gonna have AI do it. Like it doesn't quite add up the the way, you know. I think people might think it does. We we still need accountants.

Parker 33:32

Yeah, we need, we need, yeah. So I talked to a tax partner from a mid-sized accounting firm. So she said to me, she said, you know, my clients, they could have just asked ChatGPT of what to do, right? How to file for your taxes. But she said, my phone has never ran so often because they just want to get information from me, like whether what their ChatGPT says is true [laughter]. Right? So we need someone to make the final call that whether that is actually the case. And and another example, just my personal experience. I was recently writing a blog post with one of my economist friends, and she wanted to do this legal analysis. And and my father-in-law and my husband, they are lawyers. And so she's just said, hey, I used AI to help me do this analysis, but I'm really not sure whether the AI is just , you know, trying to please me and say that yes, you are correct, or I am actually correct. So I have to beg my husband, my father-in-law to say, "Hey, can you please take a look at this analysis, whether that's correct?" So that just speaks to you about why we still need those professionals to give you the actual information in this world of like availability everywhere. Right. So yeah.

Halbesleben 34:57

Yeah. That's interesting. You you mentioned accounting being sort of an industry of trust.

Parker 35:02

Yes.

Halbesleben 35:03

And then when we don't trust the AI, we we have to go back to the field and the and the individuals. That's what's really interesting.

Parker 35:09

Yeah, yeah, yeah, yeah, yeah. And also I think in accounting, actually, we need to really broaden what accounting does. So now for let's say in the audit area, right? Traditionally we think about we audit financial reports, but now we as a profession is expanding to providing assurance on the ESG reports. AI, right, that's going to be a big business. And I already talked to some people who are in the regulatory field. They are very interested in this field because now, if more firms are using AI, we as accountants, this is a business opportunity, right? We should step in and say, hey, we've been providing assurance for decades. We actually can provide the same principle, and then we can audit AI, whether your AI is running, you know, following these rules. So I think we also need to expand what we do as a profession, not just like financial reports, those traditional things.

Halbesleben 36:02

Yeah, yeah. So what's up next for you in the research space? What's something interesting that you that you're free to share?

Parker 36:09

Yeah, free to share. I think so I think I always want to do something that is really interesting, and I try to be out of the box. Something I want to do, which can be ambitious, is I want to do some like a simulation-based tasks to look at AI-enabled accounting fraud. Because now we think about AI, we always just say, hey, AI is good, blah, blah, blah. AI can improve this and that. But in reality, we see some news articles about how people just generate fake invoices, right? So there are some issues to consider that the nature of accounting fraud is changing in AI, and how do we do that? And but to publish at the top journal, I have to think about a strategy, like what kind of method should I use? Because I cannot just say blah, blah, blah.

Halbesleben 37:02

Right.

Parker 37:03

Or yeah, so I have to think about a research methodology to do it. So the only way I can think of is a simulation. Like, for example, let's say if I have a model, if I change this parameter and the manager justifies that parameter, and that can lead to a material misstatement, something like that. I just wish I had more time to do it.

Halbesleben 37:24

Yeah, yeah, goodness. Sound it sounds like really interesting work, though. I mean, the kind of work that'll have a big impact as we move forward.

Parker 37:30

I hope I just think about it and I feel, that's really interesting. But then I look at my schedule, I have to push this out, I have to do this, I have to so it's in my mind right now.

Halbesleben 37:40

Yeah. Well, you know, I along those lines, you know, from talking with you in some other settings, and you know, like you mentioned your husband, I know you're a relatively new mom. Congratulations. What outside of the research, what are some things that that interest you just as a person?

Parker 37:56

Yeah, so I think in general I'm pretty nerdy [laughter]. So since I became a mom, I think my interest is generally parenting-related stuff.

Halbesleben 38:05

Uh-huh.

Parker 38:06

When I'm not working, I try to just and if the baby is not, if I don't have to take care of the baby right now, I just try to get more workout. I get on treadmill, I get on any machine, and then I put on podcasts. The podcast is always about parenting. And just yeah, I think being a parent is amazing because I never thought having this baby will bring so much joy to my life, to my family's life, and it just feels so great to kind of grow up again with him because you are watching him to develop from this fragile baby to now sitting up and exploring and then moving on to the future. I just feel it's like very great per like for me, it's a personal growth opportunity that you are experiencing the world with him and grow with him. So I love those podcasts or lessons about the psychology behind the child rearing, different methodologies. And before I became a mom, I just in in my free time, I just like to watch TV. Yeah, I yeah, I wish I had some hobby, like for example, oh, I like to go do like a biking, whatever. I'm not that kind of person. Like to me, at most, I just walk my dogs in the in the neighborhood that satisfies me. And and I also like my husband and I love exploring new neighborhoods or new restaurants, and we just drive to a random place and say, hey, there's a new restaurant, let's try it out. And now it's tougher with the baby, but now since he can sit up, we are exploring some new places. So...

Halbesleben 39:53

That's great.

Parker 39:54

That's been exciting. Yeah.

Halbesleben 39:55

That's great. Well, I can confirm as a dad of three that it is pretty amazing to - to raise kids. I also can confirm that I don't have any hobbies either [laughter]. I just drive my kids around to practice and

Parker 40:07

Because when you get free time, you feel I just want to relax.

Halbesleben 40:11

Exactly.

Parker 40:11

Sorry.

Halbesleben 40:12

Yeah, exactly. So well, Abby, it has been such a joy to talk with you today. I'm so glad that we've got amazing researchers that are working on important issues like the work that you're interested in. And so thank you for being here today. More importantly, thank you for the great work that you're doing. And I'm so happy I had a chance to talk with you.

Parker 40:29

Yeah, I'm so glad to be here too. Thank you.

Halbesleben 40:32

Thank you. Thank you for listening to the Inside Alvarez Business Podcast. Special thanks to our producer, Brittney Johnson, and for the support of Wendy Frost and Melissa Lackey to help make this podcast possible. Stay connected with the Carlos Alvarez College of Business at the University of Texas in San Antonio to learn more about how we are empowering the next generation of business thinkers and conducting groundbreaking research to ensure their success. Follow us on social media or visit us online at business.utsa.edu. Until next time, I'm Jonathon Halbesleben. Thank you for listening.