Tech roundtable: There’s more to AI than ChatGPT

Editor’s note: Blockchain, digital assets, and cybersecurity will be covered in more depth in the May JofA.

A year after likening generative artificial intelligence (GenAI) to Wikipedia and labeling ChatGPT as immature, the participants in the JofA Accounting Technology Roundtable gathered to offer their assessment of the class of chatbots that has captivated the public, accountants included, over the past 18 months.

The 2024 roundtable, recorded in January, featured the same three experts (see the sidebar, “About the Panelists,” near the bottom of this page) who participated in the 2023 edition of a JofA annual tradition dating back to 2011. Besides discussing generative and other types of AI, the participants explored numerous other topics. The conversation covered so much ground that it was split into two episodes of the JofA Podcast. The full conversation is available here. The transcript has been edited for length and clarity and will also be presented in two parts. Topics featured in part 1 include AI, automation, and Web 3.0. Part 2, which will be published in the May JofA, explores topics that include “people tech,” blockchain, digital assets, and cybersecurity.

What is GenAI and how does it differ from other types of AI?

Wesley Hartman: The way I describe it, generative AI is really just a fantastic computer algorithm that functions like an auto-complete with data from the internet. Obviously, there’s a lot more to it, but generative AI generates its responses based on information it finds on the internet, which is still data generated by people. So, it’s a low-level AI that can find answers but can’t really think for itself. The next level is artificial general intelligence, which can create new ideas. The final level is advanced artificial intelligence or super artificial intelligence, which in some ways might become a little scary and not as distinguishable from a human. But those last two levels are theoretical. Generative AI is what we have right now.

The biggest brand name in GenAI is ChatGPT. When we talked about ChatGPT last year, it was only a few months old. And Amanda, you compared it to Wikipedia. A year later, what is your assessment?

Amanda Wilkie: I think we have to provide a little context. I did equate it to something like Wikipedia but in the sense that I wouldn’t recommend that you take anything that comes out of it as authoritative. Because ChatGPT was trained on public information, there are a lot of errors in its responses. Sometimes, it makes things up, which is referred to as a hallucination. So, you still can’t say ChatGPT is authoritative.

In the last 12 months, people have put time, effort, and a lot of capital into generative AI. We’re starting to see a lot of different uses for it. OpenAI, the owner of ChatGPT, has partnered with Microsoft to create tools where private companies can create their own GPT technologies. They have recently announced that they are creating a ChatGPT store — think the Apple App Store or Google Play — where individuals or companies that have created their own applications can make them available to consumers. What if a firm or one of the vendors in this space creates a ChatGPT to train and educate on how to do a tax return or how to properly look at financial statements?

Donny Shimamoto: I don’t know if a lot of firms realize that we’ve been using AI for years. Quick- Books has had it in its account reconciliation feature for like 10 years. Botkeeper [a company that uses AI to automate bookkeeping work] has been around for eight years. I was talking with their CEO, Enrico Palmerino, and he said they finally feel like they have maturity in their models.

And here’s one of the risks with these models, especially if you’re doing it from scratch: Enrico was saying that one of the things that they were worried about is that if you have data come in that’s not clean or accurate, it breaks the model, causes it to predict the wrong thing. This is tied back to that concept mentioned earlier of how much do we trust the things from the internet to be correct? Somebody has to police that. So, Enrico was telling me that it took them eight years to get to the point where they feel like, OK, even if some bad data comes in, it’s not going to change the model.

Wilkie: I think it is a huge point that firms really need to pay attention to their data. I fear that firms are going to say, “Let’s build some sort of generative AI and then we can use it for our clients, or we can resell it.” And if they are not paying attention to the data they are training the AI on, it’s going to be a mess.

Hartman: There was a case where they fed an art generative AI model pictures of a cat, but said it was a dog. I think they fed it like 10 or 15 photos, and it basically messed up the entire model like it could no longer identify cats versus dogs.

What skills or knowledge do accountants need to develop to assess AI models or figure out the data?

Shimamoto: This is where people like Wes and Amanda come into play. It’s not about the accounting itself. I’m trained on both sides, but I rely upon my technologists — and you need to find the right one. It’s not just any IT person who’s going to be able to do this. It needs to be someone trained in this.

As accountants, we need to understand the bigger picture of the risks. We need to understand how we do policy. A lot of this comes back to data governance, which is tied to corporate governance. When we get into data governance and the risks, cybersecurity gets tied into it as well. We need to partner with IT to really make this stuff happen.

Wilkie: We’re going to see even more of a focus on that partnership between the CPA and the technologists. It’s going to be very, very important. There’s also going to be opportunity in this space for auditing and reassuring people that the technology is doing what it is supposed to be doing, what it says it’s going to be doing.

Some of those AI models, the foundational models, you’re not going to be able to pull back the layers of that, that’s proprietary. It’s going to be hard to get in there. It’s going to be hard to understand it. So, we’re going to have to use the tools and techniques that the profession has been using for a long time. Like you said, Donny, governance, IT controls, and understanding and auditing the data that’s going in and the data and information that’s coming out. These aren’t new concepts, but we’re going to be applying them in new ways when it comes to auditing and trusting AI.

AI gets tied in with automation, but not all automation is AI. What are the best automation opportunities for CPAs and firms?

Hartman: I tend to divide automation into three tiers. The first tier is, let’s look at your existing technology because you’d be amazed at how many automation tools firms just don’t use — like setting up templates for workflows or invoicing or automatic notifications, things like that. So, let’s take a first look at what technology you’re utilizing and see if there are built-in automation tools that you can leverage to automate parts of your firm, because even something small that takes one minute can add up if you’re doing it on every single return or every single step of your audit process.

The second tier is where you’re looking at Zapier or Make.com or Pipedream — where you can build your own automations between different systems or even within systems. You can create workflows. For example, I built one myself because it was getting frustrating to take Calendly information and put it into my CRM system. I was manually doing it because I wasn’t really getting a huge number of calls. But then I got to the point where I felt like I was wasting a bunch of time, so I built a Zapier and now it’s automated.

The last tier of automation really is more of a custom software development situation where the process that you’re trying to automate is maybe too complex, has too many different systems, or doesn’t have connections on Zapier or Make.com. So, you have to do an RPA where you log in through the front end because that still has to happen.

Most of the time, accounting firms, start with your software stack. After that, look at Zapier and then after that, if you want something more customized, then you look at a developer.

Wilkie: In its purest form, automation is a computer doing something that a human would normally do. When I talk to firms, I tell them, “If you’re entering the same information in more than one system, you are creating an opportunity for someone to forget to enter it or to key it in incorrectly, making your data get out of sync.”

A simple automation is just creating that data integration so that you have one source of truth. And then that system sends that information and makes those updates to all your other systems. Again, when I talk to firms, especially small firms, that say they are frustrated because none of their systems talk to each other, that’s a great place to start for some simple automations.

Where are we on Web 3.0?

Wilkie: This time last year I was very excited about Web 3.0, and I still am, but I feel like when generative AI just kind of exploded around this time last year, a lot of people started focusing on that.

Even large companies like Apple, Meta, Google — they refocused their energy and their capital on generative AI. But there’s still some things that are happening. Apple’s actually releasing their new VR [virtual reality] headset, the Apple Vision Pro.

I don’t know if this is like the first iPhone or first iWatch, or if we’re going to need a couple of iterations for it to really take off. But with it being in the Apple ecosystem and being a smaller, more powerful headset opposed to the Oculus, I think we’re right on the edge of VR becoming more mainstream and not just for like playing games and things like that.

We talked last year about using VR for things like training and doing some meetings. We’re seeing a little bit of that, but it’s definitely not taken off like I had hoped that it would. But I think we’ll still see it.

Just to recap, Web 3.0 is that next iteration of the internet. It is much more immersive through technologies such as VR, but that means that we have to figure out things like an economy in that world. It brings in crypto or digital assets. It brings in blockchain. We’re going to have to figure out privacy and identity. That’s going to bring in identity on the blockchain. That’s going to bring in cybersecurity.

About the panelists

• Donny Shimamoto, CPA/CITP, CGMA, is founder and managing director of IntrapriseTechKnowlogies and inspiration architect of the Center for Accounting Transformation.
• Wesley Hartman is founder of robotic process automation developer Automata Practice Development, former director of technology at a midsize public accounting firm, and co-author of the JofA’s Technology Q&A column.
• Amanda Wilkie is a consultant with Boomer Consulting and former chief information officer at one of the 25 largest public accounting firms in the United States.

About the author

Jeff Drew is the JofA’s editor-inchief. To comment on this article or to suggest an idea for another article, contact him at Jeff.Drew@aicpacima.com. To comment on this article or to suggest an idea for another article, contact Jeff Drew at Jeff.Drew@aicpa-cima.com.

LEARNING RESOURCES

Unleash the Power of AI: A CPA’s Guide to ChatGPT

In this seminar, you’ll learn how to use ChatGPT to streamline your workload, save time, and boost overall efficiency.

April 19, 1 p.m. ET

WEBCAST

AICPA & CIMA ENGAGE

Technology is one of more than a half-dozen tracks delivering up to 42 CPE hours at the biggest accounting event of the year, held at the ARIA in Las Vegas.

June 3–6

CONFERENCE

For more information or to make a purchase, go to aicpa-cima.com/cpe-learning or call 888-777-7077.

AICPA & CIMA RESOURCES

Articles

JofA AI coverage page

“What AI Can Do for Auditors,” JofA, Feb. 1, 2024

“CPAs at a Crossroads: Real Approaches to Artificial Intelligence,” JofA, Jan. 8, 2024

“Gen AI, Business Model Transformation, and More With AICPA & CIMA CEO,” JofA, Jan. 4, 2024

“Getting Started With Generative AI/ChatGPT,” Professional Insights, Sept. 30, 2023

“The Promise and Peril of ChatGPT,” JofA, May 1, 2023

Podcast episode

“Why Should CPAs Understand ChatGPT?” JofA, Oct. 5, 2023