3010: The Intersection of AI and Ethics in Accounting

[00:00:01] [SPEAKER_01]: Is technology merely a tool? Or can it profoundly reshape industries and ethical boundaries?

[00:00:09] [SPEAKER_01]: Well, today here on Tech Talks Daily I'm going to be joined by Dr Mofan Afkan and he's a

[00:00:14] [SPEAKER_01]: renowned figure from Methodist University and he has an infectious enthusiasm for all

[00:00:20] [SPEAKER_01]: things technology. Especially how artificial intelligence is reshaping the field of accounting.

[00:00:27] [SPEAKER_01]: But with his extensive experience in academia and service on the audit committee in North

[00:00:34] [SPEAKER_01]: Carolina, for me he brings a unique perspective on the intersection of AI, ethics, accounting,

[00:00:42] [SPEAKER_01]: education and so much more. Honestly, he's the kind of guy I could listen to for hours

[00:00:47] [SPEAKER_01]: and so much so I had to stop myself going down a rabbit hole at an additional side

[00:00:52] [SPEAKER_01]: mission as we were talking today so I will be getting him back on later in the

[00:00:55] [SPEAKER_01]: year. But no more spoilers from me.

[00:00:58] [SPEAKER_01]: Delivering daily content to 140,000 of you wonderful monthly listeners across the globe is no small

[00:01:04] [SPEAKER_01]: feat. I don't want to take all the credit here because it wouldn't be possible without the backing

[00:01:08] [SPEAKER_01]: of our dedicated sponsors and partners. And I also want to shine a light on the fact that legacy

[00:01:13] [SPEAKER_01]: managed file transfer tools are looking dated. They often lack the security that today's remote

[00:01:18] [SPEAKER_01]: workforce demands and companies that continue relying on that outdated tech, they're in

[00:01:25] [SPEAKER_01]: IT professionals. Are you tired of juggling these multiple servers for secure file sharing,

[00:01:31] [SPEAKER_01]: integrated shared folders and email plus a comprehensive rest API?

[00:01:36] [SPEAKER_01]: Kiteworks simplifies your workflow. For administrators, you can experience unmatched

[00:01:40] [SPEAKER_01]: functionality and integration that traditional MFT servers just can't touch.

[00:01:45] [SPEAKER_01]: Step into the future of secure managed file transfer with Kiteworks by going to

[00:01:50] [SPEAKER_01]: kiteworks.com to get started. That's kiteworks.com and remember Kiteworks is also FedRAMP

[00:01:56] [SPEAKER_01]: moderate authorised. Thank you for your patience today, this is the moment you've been waiting

[00:02:00] [SPEAKER_01]: for. It's time to welcome my guest onto the show. Buckle up and hold on tight because no matter

[00:02:06] [SPEAKER_01]: where you are in the world right now, it's time for me to beam your ears all the way to

[00:02:10] [SPEAKER_01]: North Carolina where my guest is waiting to talk with us. So a massive warm welcome to the

[00:02:18] [SPEAKER_01]: show. Kelly Tidewell listening a little about who you are and what you do.

[00:02:23] [SPEAKER_00]: Well, first Neil, I'm excited to be here. My name is Dr. M. Fodokpun. I'm assistant professor

[00:02:29] [SPEAKER_00]: of accounting at Methodist University in Fayetteville, North Carolina. I teach

[00:02:36] [SPEAKER_00]: graduate and undergraduate level courses in accounting. I am an AI and metaverse enthusiast.

[00:02:46] [SPEAKER_00]: So I love technology. I think a better way of saying that is I really love technology.

[00:02:51] [SPEAKER_01]: And I love that. That's one of the reasons I invite you on here. And what I try and do

[00:02:55] [SPEAKER_01]: every single day on this podcast is talk about something that people are talking about in

[00:03:00] [SPEAKER_01]: businesses and maybe demystify some of the complex technologies and maybe ease a few concerns.

[00:03:06] [SPEAKER_01]: And right now everyone's talking about AI, but the reason I invited you on today was to talk

[00:03:11] [SPEAKER_01]: about AI ethics, which is another huge talking point right now. So to begin with and set the

[00:03:17] [SPEAKER_01]: scene for our conversation today, can you just walk me through the key principles behind the

[00:03:22] [SPEAKER_01]: accounting framework for AI ethics? And also how that aligns with the AI code of professional

[00:03:30] [SPEAKER_01]: conduct as well. There's a lot going on there, but just for anyone that's hearing about this

[00:03:35] [SPEAKER_01]: subject for the first time, could you just set the scene for me? I think that is a huge

[00:03:41] [SPEAKER_00]: area that is still being worked on as far as AI ethics. And tying that in to the professional

[00:03:51] [SPEAKER_00]: code of conduct, so the AI CPA code of conduct, there's still discussion on that.

[00:03:56] [SPEAKER_00]: There's still challenges with that. And I think as the technology continues to develop,

[00:04:03] [SPEAKER_00]: there'll be even more and more challenges that we'll see as far as how AI is used, privacy issues,

[00:04:13] [SPEAKER_00]: it's particularly data. I think that's a big issue. What area that I'm interested in is

[00:04:21] [SPEAKER_00]: devices. So we've got so many of these AI powered devices. I've had conversations myself

[00:04:30] [SPEAKER_00]: with a colleague, which this may be interesting. So I've got, if you're familiar with the Rayband,

[00:04:39] [SPEAKER_00]: the Meta Rayband glasses, AI glasses, I've got a pair of those. So I had a conversation with

[00:04:48] [SPEAKER_00]: my colleague. So I'm like, well, how would you, if I was working at a company and I had

[00:04:55] [SPEAKER_00]: a pair of these glasses, number one, I could come into the office. I could tap, tap on my glasses

[00:05:04] [SPEAKER_00]: and take pictures stuff on my screen. I could go with a client, take pictures of a client. I can

[00:05:10] [SPEAKER_00]: record parts of the conversation. I can take pictures of documents. So there's a lot of

[00:05:16] [SPEAKER_00]: things that I could do with these glasses. And no one would know. So the question was, well,

[00:05:23] [SPEAKER_00]: what policy would this fall under? And my colleague was like, I believe it will fall under a smartphone

[00:05:30] [SPEAKER_00]: policy, which I thought was interesting. So the way we have a smartphone about how you use your data,

[00:05:36] [SPEAKER_00]: how you're not supposed to, but again, how would you know if I have my phone,

[00:05:42] [SPEAKER_00]: you know that that's a phone. But if I have these glasses on, unless I tell you,

[00:05:48] [SPEAKER_00]: you really wouldn't know. So I think that you could say that's a challenge. It's not a challenge.

[00:05:56] [SPEAKER_00]: You can say, well, okay, we can use smartphone policy. But does that really cover something

[00:06:01] [SPEAKER_00]: like that with these devices? And then when I thought about these the rabbit pin and the

[00:06:06] [SPEAKER_00]: human, you know, when you see these different pins and these other device applaud is another

[00:06:11] [SPEAKER_00]: one where you can put it on your phone and it records everything. So, you know, I start to

[00:06:17] [SPEAKER_00]: think about how, how are we going to mitigate those challenges? And especially from what is right,

[00:06:23] [SPEAKER_00]: what is wrong? Is it okay for me to record these things with my applaud device or take pictures

[00:06:29] [SPEAKER_00]: of documents with my meta AI glasses? And does is that considered a smartphone? Or if I'm in

[00:06:37] [SPEAKER_00]: trouble because I'm using it, can I say, well, it's not a phone. These are my glasses. So

[00:06:43] [SPEAKER_00]: so I, you know, I think there's a lot of challenges. And I think these challenges with AI and this is

[00:06:51] [SPEAKER_00]: all stemming from AI are coming at us pretty quickly. And the fact that we're having these

[00:06:56] [SPEAKER_00]: conversations, I think there is something there. You know, it's not cut and dry.

[00:07:03] [SPEAKER_01]: And I'm glad you raised that. There'll be a few people who probably don't know about

[00:07:06] [SPEAKER_01]: the Ray-Ban glasses. Obviously, we know about the Vision Pro, that big bulky headset. There's

[00:07:13] [SPEAKER_01]: Mixed Reality headsets that are big and bulky. But these Ray-Bans look just like a pair of glasses.

[00:07:19] [SPEAKER_01]: Are they like the Google glasses is a little red light? Come on, when you're recording something or

[00:07:23] [SPEAKER_01]: if I was sat next to you in a bar or a restaurant, would I know that you were recording with these

[00:07:29] [SPEAKER_00]: glasses? These are my glasses. I don't know if they'll be able to see, but they have a little

[00:07:36] [SPEAKER_00]: on the side of them. They have a little camera on the side, but pretty much you

[00:07:41] [SPEAKER_01]: wouldn't be able to know. What's been your experience with those just out of interest?

[00:07:49] [SPEAKER_00]: They've been great. I mean, they work better than expected.

[00:07:54] [SPEAKER_00]: You're able to talk to the AI so you can ask it questions. What is that? If you see something,

[00:08:00] [SPEAKER_00]: what is this? So you can ask Vision so he can read things. I mean, it's a very

[00:08:07] [SPEAKER_00]: interesting technology. So if you take a picture, it goes to your phone.

[00:08:13] [SPEAKER_00]: So then you have it on your phone and then you can post it to social media or save it in your phone

[00:08:19] [SPEAKER_00]: and you have those documents or you have those images.

[00:08:23] [SPEAKER_01]: Wow, that's incredible. I think we could almost dedicate an entire episode to that because I

[00:08:28] [SPEAKER_01]: know you are passionate about that area. It sounds like we're going to need to get

[00:08:31] [SPEAKER_01]: you back on there. But if we go back to the world of accounting for a moment, what would you say are

[00:08:37] [SPEAKER_01]: the most significant ethical considerations when integrating AI technologies into the account

[00:08:43] [SPEAKER_01]: in sector? Because again, there's a lot of businesses now they're rushing in, they're

[00:08:47] [SPEAKER_01]: getting the new AI stuff. But what kind of things should they be thinking about, especially from

[00:08:54] [SPEAKER_00]: standpoint? One thing we hear privacy. So I think that's one big thing also outside use.

[00:09:04] [SPEAKER_00]: So there's a statistic where you have, I think was it 75% of knowledge workers are using AI,

[00:09:15] [SPEAKER_00]: but they're using it not at work. So I think that could be a concern

[00:09:22] [SPEAKER_00]: where you may not either you have the tools at the office or you don't have it at the office,

[00:09:30] [SPEAKER_00]: but then you have employees who are using it because they want to be more productive.

[00:09:34] [SPEAKER_00]: They want to get more done. So they're using this on their own computers, which may not have

[00:09:40] [SPEAKER_00]: the same privacy protocols if you have an enterprise license for some of these models.

[00:09:47] [SPEAKER_00]: So it ties into privacy, people using it, I guess off the books. I don't know if that's a way to say

[00:09:53] [SPEAKER_00]: it. I think that could be a big concern as well because you don't know how they're using it and

[00:09:59] [SPEAKER_00]: it's not managed. They're using it outside of the workplace. So you can't really manage how

[00:10:06] [SPEAKER_00]: they're using it or understand how they're integrating it into their work processes.

[00:10:13] [SPEAKER_01]: So if we've got an accountant listening to our conversation today, anywhere in the world, maybe

[00:10:18] [SPEAKER_01]: they're sat on the fence. Maybe one side of them is thinking, oh, it's all overblown. It's overhyped

[00:10:23] [SPEAKER_01]: or it's set for the future or I'm too scared. I want nothing to do without ethical reasons.

[00:10:28] [SPEAKER_01]: Can you just set the scene on how AI is transforming the account profession in terms

[00:10:34] [SPEAKER_01]: of things like productivity and accuracy with some real-world use cases and also on

[00:10:39] [SPEAKER_01]: the flip side, just some of those potential risks that need to be managed to ensure ethical AI

[00:10:45] [SPEAKER_01]: adoption. I understand there's a lot going on in there as well but can you share around that?

[00:10:51] [SPEAKER_00]: That was a big question and it was a good question. The first part when you said it's

[00:10:58] [SPEAKER_00]: overblown, I would argue it's underblown. It's the opposite.

[00:11:06] [SPEAKER_00]: Why would I say that? From my perspective and I get it, we tend to focus on the negative side.

[00:11:17] [SPEAKER_00]: So we focus on hallucinations, meaning the AI lies. In other words, it hallucinates.

[00:11:26] [SPEAKER_00]: It's not very reliable. You ask it nine questions, it'll get those nine questions right,

[00:11:32] [SPEAKER_00]: but that tenth one, it'll get wrong. It's not like a calculator. You sit with your calculator,

[00:11:39] [SPEAKER_00]: you can type in 2 plus 2 equals 4 10 times and you'll get the same answer.

[00:11:47] [SPEAKER_00]: So we don't have that reliability. Yes, issues with privacy. That's there.

[00:11:53] [SPEAKER_00]: But here's the flip side. If you think about the capabilities,

[00:12:00] [SPEAKER_00]: so whenever I look at these assessments, so if you go right now to Anthropics website

[00:12:08] [SPEAKER_00]: and you look at the performance for, there's a Claude 3.5 sign and you look at that,

[00:12:15] [SPEAKER_00]: it'll say for the undergraduate assessment, the benchmark scores 86%.

[00:12:22] [SPEAKER_00]: So he gets 86% of these questions right. I think chat GPT is about 86% as well. Then the master's

[00:12:31] [SPEAKER_00]: level assessment is around 50% for Claude and I think 30% for a chat GPT. Why is that interesting?

[00:12:43] [SPEAKER_00]: And why is that underblown? Well, if you look at education attainment data in the US,

[00:12:54] [SPEAKER_00]: about 40% of Americans have a bachelor's degree. So that's below, that's not even 50%.

[00:13:01] [SPEAKER_00]: You look at graduate degrees, somewhere 13 to 15%.

[00:13:07] [SPEAKER_00]: Right. And then you think about a doctorate is somewhere around 3%, 5% there.

[00:13:15] [SPEAKER_00]: So when you hear them talk about these models, we want to get it to be better

[00:13:20] [SPEAKER_00]: to work at a PhD level. I'm saying to myself, well, it's already working

[00:13:27] [SPEAKER_00]: at an undergraduate level, which is at a level higher than most Americans.

[00:13:33] [SPEAKER_00]: Yeah. So when they talk, you see what I'm so when you talk about AGI, you talk about how powerful it is.

[00:13:41] [SPEAKER_00]: The other thing is you think about the literacy rate, literacy rate in the US about 50 something,

[00:13:49] [SPEAKER_00]: 54, I think it's 54% of Americans read at a sixth grade level. You've got a tool

[00:13:56] [SPEAKER_00]: that scores 86% at an undergraduate level. So I think this when they talk about AGI being

[00:14:03] [SPEAKER_00]: smarter than the average human, Ali is already smarter than the average person in my opinion.

[00:14:08] [SPEAKER_00]: That may be controversial, but if you kind of compare where are normal people because many

[00:14:15] [SPEAKER_00]: you see them talking, I hear Mira, Marati, PhD, PhD, PhD, 3% in at least in the US are PhDs.

[00:14:26] [SPEAKER_00]: And then you've got to peel back the onion on that too. My doctorate's in a county.

[00:14:33] [SPEAKER_00]: I couldn't if I went to a large language model and asked a PhD question in physics,

[00:14:40] [SPEAKER_00]: I would know if it's right or wrong or what it is. I'd have no idea. So I think when you talk about

[00:14:48] [SPEAKER_00]: it's being over high, I don't think it's being over. I think it's actually the opposite.

[00:14:52] [SPEAKER_00]: This they're not really talking about how powerful and capable the models are that we have right

[00:14:59] [SPEAKER_00]: now. Now to the flip side about the issues when we talk about use and ethics in workflow,

[00:15:10] [SPEAKER_00]: that's a challenge because you have to do so there's all these questions in accounting.

[00:15:16] [SPEAKER_00]: If I'm working on something and do I tell my client I'm using AI to do this or do I tell them

[00:15:23] [SPEAKER_00]: that I'm using it, that's an ethical issue. They're paying me to do it but hey, I'm using AI to do it.

[00:15:31] [SPEAKER_00]: Do I tell them that hey, about 80% of this was done by AI and 20% was done? So those questions

[00:15:40] [SPEAKER_00]: can come up and I think that can be a concern as well as the privacy issue as well.

[00:15:48] [SPEAKER_01]: 100% with you and I also agree with you when you said we do focus on the negative too much sometimes.

[00:15:55] [SPEAKER_01]: Another reason I invite you on today was a paper that you co-authored also discusses

[00:16:01] [SPEAKER_01]: the transformative potential and challenges of AI in accounting education but can you tell me a

[00:16:07] [SPEAKER_01]: bit more about how AI can enhance learning experiences and prepare students for the

[00:16:12] [SPEAKER_01]: future workforce because this isn't something we should be celebrating, right?

[00:16:17] [SPEAKER_00]: Yeah, we should because they're using it. Everybody, I mean when you think about the large,

[00:16:22] [SPEAKER_00]: so you think about the big four firms, they're heavily investing billions of dollars in the AI

[00:16:29] [SPEAKER_00]: partnering, implementing, adding it to workflows because they want to increase productivity.

[00:16:35] [SPEAKER_00]: They want to get things done faster and over and over again I hear from my colleagues at

[00:16:42] [SPEAKER_00]: these firms generate and review, generate and review. So what that means is that

[00:16:51] [SPEAKER_00]: you're looking for ways to add it to the workflow. Okay, if there's some sort of process

[00:16:58] [SPEAKER_00]: that would take me six hours to do it. Now, if I use AI to generate it in maybe,

[00:17:07] [SPEAKER_00]: it may take 10 seconds to generate whatever it is I need and then I spend two hours reviewing,

[00:17:16] [SPEAKER_00]: I've saved about four hours, right? So that is the main, there's other use

[00:17:26] [SPEAKER_00]: came but that's the main one. Generate and review, increasing that productivity

[00:17:32] [SPEAKER_00]: and getting things done a lot faster than you would before. So then you can move on to other work,

[00:17:39] [SPEAKER_00]: you've got more capacity and as the lotto's get better and when I mean better as far as reliability.

[00:17:48] [SPEAKER_00]: So when you think about it from that generate and review, a paradigm,

[00:17:54] [SPEAKER_00]: eventually the models are going to get better and better, the review time is going to get

[00:17:59] [SPEAKER_00]: shorter and shorter and shorter because you're going to become more precise. So it'll only enhance

[00:18:05] [SPEAKER_00]: your capability. Then by adding it to workflows, you start to understand because these models for the

[00:18:13] [SPEAKER_00]: lowest part are general. So you can add it here, you can try to figure out how I can use it here,

[00:18:19] [SPEAKER_00]: maybe I can save 20 minutes here in this process. So you look for different ways to

[00:18:24] [SPEAKER_00]: use it and implement it and that productivity adds up in the classroom having students being exposed to

[00:18:32] [SPEAKER_00]: it and using it. It helps number one, they'll understand how it works, the good side, the bad

[00:18:39] [SPEAKER_00]: side to it, but also how it can be implemented and put into workflows because this is new. So

[00:18:47] [SPEAKER_00]: new technology, but it develops their skills with the technology so they're not so that

[00:18:53] [SPEAKER_00]: they're not just blind, blind or starting at zero with.

[00:18:59] [SPEAKER_01]: And something else we need to highlight that this stuff is here right now. It's not something for the

[00:19:04] [SPEAKER_01]: future. It's here right now as you mentioned a few moments ago and how smart it is right now. So

[00:19:10] [SPEAKER_01]: how important is it for educators, policymakers and industry leaders to collaborate right

[00:19:16] [SPEAKER_01]: now as well in preparing students for the evolving technology landscape? And

[00:19:21] [SPEAKER_01]: what steps can they take now to foster this collaboration? Anything that you're seeing here?

[00:19:27] [SPEAKER_00]: Yeah, to foster implementing the technology, making it very accessible because it's very important

[00:19:34] [SPEAKER_00]: that that's a good point. It is here right now. People are using it and to put yourself in the

[00:19:43] [SPEAKER_00]: sh-. And so the other thing is, I'm a professor, I'm going to call it students. I've got to

[00:19:48] [SPEAKER_00]: prepare them. If I got a freshman, I've got to prepare them for the world four years from now.

[00:19:54] [SPEAKER_00]: And we're using it here now, but it's going to be here even more and we're going to be using

[00:19:59] [SPEAKER_00]: it at a different pace. So I've got to try to expose them to as much as possible, give them

[00:20:06] [SPEAKER_00]: as much exposure and skills as possible now so that they can have a leg up when they get out

[00:20:12] [SPEAKER_00]: to the workforce. When I always use this analogy, say you're an employer, you got four candidates,

[00:20:21] [SPEAKER_00]: right? You got four candidates. Let's say they're all mathematicians, same education,

[00:20:27] [SPEAKER_00]: everything is the same except for their tools that they use to do their work,

[00:20:35] [SPEAKER_00]: their mathematics, their calculations. You got one mathematician. He only uses a pencil and paper

[00:20:43] [SPEAKER_00]: to do his calculations. So he'll be there, he'll get the job done, but it might take him two days

[00:20:52] [SPEAKER_00]: to get it done. Now you've got the other mathematician, he comes in, he's got the calculator.

[00:21:00] [SPEAKER_00]: He does all these calculations on this calculator. He can get the work done in one day.

[00:21:06] [SPEAKER_00]: Now you've got another guy comes in, he brings in his laptop computer.

[00:21:10] [SPEAKER_00]: He's got the internet and spreadsheets. He's got everything there. He can get everything done

[00:21:15] [SPEAKER_00]: in about six hours. Now you've got your other one. He's got a computer and he's got AI

[00:21:23] [SPEAKER_00]: and he can get all of these work done in one hour. Now if you're an employer, businessman

[00:21:29] [SPEAKER_00]: and you want that mathematician, which one would you want to hire?

[00:21:35] [SPEAKER_00]: Which one would you want the guy with the pencil and paper? No, you want the guy with the AI.

[00:21:42] [SPEAKER_00]: And that's where we are right now. So when we're talking about education, we got to think about

[00:21:48] [SPEAKER_00]: it this way. Employers, hey, they want you to execute. They want the output. They're going

[00:21:56] [SPEAKER_00]: to say, well, that's great. He can do all of this on pencil and paper, but it takes him 10 hours

[00:22:03] [SPEAKER_00]: to do it. I'll let the guy get it done in 15 minutes. So that's where we are right now.

[00:22:10] [SPEAKER_00]: And that's only going to in-cruits. So when I use that analogy, you kind of think about, well,

[00:22:17] [SPEAKER_00]: okay, that makes sense. Okay, I'll understand that. Why this is important and why it's

[00:22:23] [SPEAKER_00]: important for students to learn about it. And you hear me say, get exposed to it.

[00:22:31] [SPEAKER_00]: I say get exposed because it's changing so fast. So by the time you, you know,

[00:22:38] [SPEAKER_00]: I had an activity in my class last fall using Claude II. Now we've got Claude III.5.

[00:22:45] [SPEAKER_00]: By the time we start classes again, I don't know what we'll have. Right? So you got to try

[00:22:51] [SPEAKER_00]: to expose them to as much as possible. So they're familiar with it, the general basis,

[00:22:59] [SPEAKER_00]: not necessarily diving too deep because it's going to keep changing. The goalpost is going to keep

[00:23:03] [SPEAKER_01]: moving. You raise such a great point now because when I scroll down my news feed, I do see a lot of

[00:23:09] [SPEAKER_01]: education institutions that are saying, hey, we need to get everyone's smartphone, put them in

[00:23:14] [SPEAKER_01]: a pouch, lock them away. It's class time and AI, if we catch you using AI, we're going to

[00:23:20] [SPEAKER_01]: punish you for that as well. There's almost kind of two worlds going on. I understand the reasons

[00:23:25] [SPEAKER_01]: for it and the distractions, et cetera, and learning for yourself. But when these students

[00:23:30] [SPEAKER_01]: leave the education, they're entering a world where everyone's got all of these tools and

[00:23:36] [SPEAKER_01]: relying on them. So is it better to give them access and teach them how to use it responsibly?

[00:23:42] [SPEAKER_01]: It's such a big topic. Where do you stand on that?

[00:23:45] [SPEAKER_00]: It's better to give them access and have them use it responsibly because they're going to be in the

[00:23:51] [SPEAKER_00]: workplace. They're not going to be sitting in the classroom and we're preparing them for work.

[00:23:57] [SPEAKER_00]: So I think definitely getting them exposed to as many tools as possible only helps them

[00:24:05] [SPEAKER_00]: when they leave, because employers, again, they're going to be looking for the technology

[00:24:13] [SPEAKER_00]: skills. They're going to be looking for these specific skills when they get out there and

[00:24:19] [SPEAKER_00]: they want them to be able to use them, understand them, and have some level of proficiency. It

[00:24:24] [SPEAKER_00]: doesn't have to be perfect but to have some sort of proficiency with the tools. So definitely,

[00:24:30] [SPEAKER_00]: particularly in a county. It's here. And when you think about the big four firms,

[00:24:37] [SPEAKER_00]: okay, they're heavily invested. All of this trickles down. So eventually you've got other firms,

[00:24:45] [SPEAKER_00]: so you've got your regional, your smaller firms, they're going to be invested in it and using it

[00:24:51] [SPEAKER_01]: as well. Yeah. And even when you leave school, it's not a case of then getting a job and then

[00:24:57] [SPEAKER_01]: sticking in that job for 40 years and getting a wristwatch at the end of it. We're all on this

[00:25:02] [SPEAKER_01]: path of continuous learning now. So as AI continues to evolve, where do you see lifelong learning or

[00:25:09] [SPEAKER_01]: what role do you see lifelong playing in helping accounting professionals? I think indeed anyone

[00:25:14] [SPEAKER_01]: staying current with these emerging technologies and ethical standards and any tips on how they can

[00:25:20] [SPEAKER_01]: do that because there's that much information coming out as the speed of change is going at

[00:25:25] [SPEAKER_01]: break next speed but at the same time, there's almost a realization that he's not going to

[00:25:29] [SPEAKER_00]: move this slow again. It won't. So there's a theory, they call it the 10-10 rule and it's

[00:25:37] [SPEAKER_00]: in one of my favorite books. It's called Where Good Ideas Come From by Stephen Johnson.

[00:25:42] [SPEAKER_00]: And they talk about this 10-10 rule. So and this is before it would take 10 years to develop

[00:25:51] [SPEAKER_00]: a technology and then 10 years for it to be widely adopted. So let's say the color TV.

[00:25:57] [SPEAKER_00]: Took 10 years to develop it and then 10 years for it to be widely adopted. You think about DVDs

[00:26:05] [SPEAKER_00]: and CDs and et cetera. So you think about if you put that 10-10, that's 20 years as a generation.

[00:26:14] [SPEAKER_00]: So there's almost, if you say 20 years of generation, that doesn't exist anymore.

[00:26:18] [SPEAKER_00]: And it gets to the point where you have to really be, to some point, technology

[00:26:26] [SPEAKER_00]: agnostic or platform agnostic. Meaning not just marrying yourself. It's important to learn

[00:26:37] [SPEAKER_00]: one platform but learn what you need on that platform that you can use on another platform.

[00:26:43] [SPEAKER_00]: What do I mean by that? Well, if you think about Gemini, you think about Claude,

[00:26:49] [SPEAKER_00]: you think about ChatGPT. You think about all of these platforms. What do they have in common?

[00:26:55] [SPEAKER_00]: You input with prompting. Okay, I need to learn prompting. So you get good at prompting

[00:27:02] [SPEAKER_00]: at the different types of prompting. There's zero shot, few shot prompting, chain of thought,

[00:27:10] [SPEAKER_00]: tree of thought, nova. So you start learning prompting because now you can use those prompts

[00:27:16] [SPEAKER_00]: on any of those platforms. So you have to start thinking about, well, okay, I could stick with

[00:27:24] [SPEAKER_00]: one. I like one but what skills can I learn that I could carry across the platforms?

[00:27:30] [SPEAKER_00]: What skills can I get good at even if these things keep changing which they do? Just like

[00:27:35] [SPEAKER_00]: I gave the example. Last fall we were using Claude 2.0, now there's 3.5. But the through line with that

[00:27:45] [SPEAKER_00]: is prompting. So you still have that prompting piece of getting good at the prompting and

[00:27:50] [SPEAKER_00]: understanding how to build skills across these various platforms. I think that's very important.

[00:27:57] [SPEAKER_01]: Completely agree with you. And we've talked about the private sector and consumers. They're

[00:28:03] [SPEAKER_01]: racing ahead with this stuff and getting to grips with it. But we also mentioned a little

[00:28:07] [SPEAKER_01]: bit about the public sector and how they could be slow to catch up or see the value in it.

[00:28:11] [SPEAKER_01]: We talked about education but in the world of public sector accounting, how do you see AI

[00:28:18] [SPEAKER_01]: impacting that area? And are there any ethical considerations or anything that's relevant

[00:28:24] [SPEAKER_01]: to them that might be listening right now? It might feel just overwhelming but as we've said

[00:28:28] [SPEAKER_01]: multiple times, he's here right now. We need to deal with it.

[00:28:32] [SPEAKER_00]: I think one issue is that when you look at the data as far as who's using it, it can be very

[00:28:40] [SPEAKER_00]: confusing. And to tell you, I don't think we really know how many people are using it,

[00:28:47] [SPEAKER_00]: who's using it and exactly how they're using it. Why is that? Because to that point,

[00:28:55] [SPEAKER_00]: you have many people using it at home and they may not use it at work and they may not let their

[00:29:02] [SPEAKER_00]: colleagues or their employer know that, hey, I'm using AI to get a lot of more work done.

[00:29:09] [SPEAKER_00]: So and because many of these tools, you can use them at a certain capacity for free.

[00:29:16] [SPEAKER_00]: So it's not paid subscription. So I think as far as people using the tools,

[00:29:22] [SPEAKER_00]: they're more in my opinion, this is just my opinion. I think there's more people using them

[00:29:28] [SPEAKER_00]: than we know. However, they're not using them directly at the workplace.

[00:29:35] [SPEAKER_00]: And then at the workplace, if your company has not embraced or invested in AI,

[00:29:45] [SPEAKER_00]: it may be challenging for them to figure out how to implement it, how to manage it,

[00:29:50] [SPEAKER_00]: how to tackle these ethical issues around it. So they may push away from it.

[00:29:58] [SPEAKER_00]: So I think that we're in this gray area right now where you've got people using it,

[00:30:05] [SPEAKER_00]: they're not using it, there's benefit and there's not benefit. Some workplaces are using it,

[00:30:10] [SPEAKER_00]: some are not. You imagine if you have your own shop, your own practice,

[00:30:17] [SPEAKER_00]: do I tell my clients I'm using it? What will happen if my clients know? Will they want,

[00:30:22] [SPEAKER_00]: do I have to, will I lose clients? Do I charge them less money? Do I let them know these things?

[00:30:30] [SPEAKER_00]: Even though offshoring, so you think about a tax practice, many people offshore that their

[00:30:37] [SPEAKER_00]: preparation, right, they hire people overseas to do a lot of their work and then they reveal

[00:30:43] [SPEAKER_00]: it. Is that the same thing? Or do people see that in the same thing? So there's all these different

[00:30:50] [SPEAKER_00]: gray areas that we have with this that has to be sorted out.

[00:30:55] [SPEAKER_01]: Another area I'd love to ask you about is around accessibility, because I think in recent years

[00:31:00] [SPEAKER_01]: we've seen a lot of authoritative content that's slowly but surely been locked away behind

[00:31:06] [SPEAKER_01]: paywalls and that could be anything from science papers to, I don't know,

[00:31:10] [SPEAKER_01]: time magazine or a newspaper or something. It's all locked behind paywalls, only

[00:31:13] [SPEAKER_01]: certain people can access information, which means everyone goes then to

[00:31:19] [SPEAKER_01]: information that's not verified maybe. Maybe that's got something to do with the misinformation

[00:31:22] [SPEAKER_01]: we see so much of, I'm not sure. But if we look to the future, if we've got to have

[00:31:28] [SPEAKER_01]: a members of society that don't have access to $20 a month for chat GPT or Claude, etc.,

[00:31:35] [SPEAKER_01]: certain members of society get left behind because this information almost helps them be

[00:31:41] [SPEAKER_01]: limitless and improves their life. They're locked out from that world. Is there a digital divide?

[00:31:47] [SPEAKER_00]: Do you see anything around that? Yeah, we have it now because you just mentioned it. There's

[00:31:52] [SPEAKER_00]: a paywall and it's throttle. So I pay for the enterprise, I have the enterprise

[00:32:00] [SPEAKER_00]: chat GPT-40. So I can use it as much as I could sit on there most of the day and use it.

[00:32:06] [SPEAKER_00]: Someone that does it, they're not able to do that. If I go to Claude 3.5,

[00:32:13] [SPEAKER_00]: I get in there and I'm doing work in there. Eventually he's going to tell me you've used

[00:32:18] [SPEAKER_00]: up all your time. You got to come back in four hours or five hours. So it's there. And I think

[00:32:25] [SPEAKER_00]: as this technology continues to scale and improve, I do believe we're going to see more and more

[00:32:35] [SPEAKER_00]: of a divide. There is a divide. We're going to see a greater divide because of the network

[00:32:40] [SPEAKER_00]: effect. You're going to have more. Eventually you'll have more and more people using these tools.

[00:32:46] [SPEAKER_00]: It's not as if in 2023 to 2024, people stop using, they are signing up for these to more and more

[00:32:58] [SPEAKER_00]: people are. And as that increases, that divide is going to increase because you'll have more people

[00:33:03] [SPEAKER_00]: using it for free. And then you'll have those people who are behind that paywall

[00:33:09] [SPEAKER_00]: and have more access, which gives them access to more productivity, more intelligence.

[00:33:16] [SPEAKER_00]: So now there's a digital divide around intelligence. And I think that goes back to what I was

[00:33:22] [SPEAKER_00]: mentioning before. When you look at these education attainment levels, you look at the

[00:33:28] [SPEAKER_00]: performance levels. Now you're going to have individuals who are going to have access

[00:33:32] [SPEAKER_00]: to a lot of capable information, intelligence and others who don't.