ChatGPT, the wildly popular artificial intelligence tool, holds significant promise for corporate finance teams, but also comes with some risks, according to Glenn Hopper, a finance tech expert and a director at Scottsdale, Ariz-based CFO advisory firm Eventus Advisory Group.

The technology, developed by Microsoft-backed startup OpenAI, is capable of interacting in a conversational way and producing responses to questions across a number of subject areas.

While interest in the technology is rapidly growing, including in the finance space, it might be a while before CFOs can fully embrace the tool due to concerns such as accuracy and data privacy, Hopper said in an interview with CFO Dive. 

OpenAI recently tightened its data-handling policies in an effort to alleviate privacy concerns.

Hopper regularly advises clients on finance tech issues. He’s also the author of a book titled, “Deep Finance: Corporate Finance in the Information Age,” as well as a paper titled, “Application of ChatGPT to Build an FP&A Tool.”

The following is a Q&A between Hopper and CFO Dive’s Alexei Alexis. Remarks have been edited for clarity. 

CFO Dive: In your book, you write: “If you’re not keeping up with technology and you’re not keeping up with the latest tools out there, you’re doing a disservice to your company.” Why is this principle so important, and does it apply to ChatGPT?

Glenn Hopper: So, in general, I've been on this mission for a number of years to help finance departments understand the technological tools that are out there. They’re getting more and more powerful, and if you're not jumping in and using these tools, then you're not getting the kind of data that can help you be more effective. And I think this does extend to ChatGPT. 

But there’s a caveat: With ChatGPT, there’s the question of how confident you can be in the data. The technology has been known to hallucinate and state things very categorically that are undeniably wrong. So, if you’re just letting it answer all of your business questions for you, that could be a problem.

CFO Dive: What are some examples of how this technology can help finance teams do their jobs better? 

Glenn Hopper: We're only scratching the surface in terms of how the technology will help us. Right now, AI feels very much like when we saw the introduction of the internet. We know we have this massively powerful thing right now, and we can't even imagine yet all that will come out of it.

In the finance space, I think you will see people incorporating the chatbot functionality into their analysis. And I think we're very close to having the chatbots being able to answer questions and even generating reports like a junior financial analyst would do.

CFO Dive: What’s your view of the costs of using ChatGPT versus the potential return-on-investment?

Glenn Hopper: Currently, the cost can be very minimal, but I don’t know what the pricing will look like in the future.

I’m always cautious when I talk about ROI on technology, because I think of it not necessarily as a cost savings, but as an investment. With AI, yes, we’re able to do work with a reduced headcount. But that’s the short-sided view. The more long-range view is that we’re changing the nature of the work we do, moving away from mindless tasks to more mindful tasks that add value. But this also requires significantly more skilled people. So, we’re making some things easier, but then the people we are hiring will need to have a higher knowledge base and skill set.

CFO Dive: For CFOs who may be trying to figure out the best way to prioritize their technology spending during a time of economic uncertainty, how much should they be focusing on ChatGPT adoption?

Glenn Hopper: The technology is a novelty right now, and it's interesting, and it gives some interesting responses, but just going to the public ChatGPT website with questions isn't going to give you much value from a finance perspective. The real value comes when you have data scientists and engineers to go along with it, or the resources to make it a part of your finance tech stack. Then, it starts to get real interesting. This is particularly possible for larger companies.

I work a lot in the SME (small and medium-sized enterprises) space. Companies of that size aren’t realistically going to have the resources to be able to build out their own tool to do this and to really integrate and incorporate the functionality of ChatGPT into their finance tech stack. But my expectation is that we’re going to eventually have an off-the-shelf product that allows you to train it to your company’s financials, to the metrics that are important to you. 

CFO Dive: Companies such as JPMorgan Chase and Verizon have reportedly barred their employees from using ChatGPT at work over concerns about the risk of compromising sensitive business data. Is this a mistake?

Glenn Hopper: No, it's not. They cited concerns about information they don't want to get out. If you think about it, ChatGPT is a massive platform. The challenge of keeping user data private and making sure that it’s not leaked out is going to be a significant issue, just like with any other kind of internet privacy situation.

CFO Dive: What advice do you have for CFOs who are potentially interested in using ChatGPT but are worried about risks such as compromising sensitive data or getting results that are flawed or inaccurate at times?

Glenn Hopper: I think the publicly available ChatGPT is, right now, a novelty and interesting, and you could use it to help you come up with ideas for, you know, your board presentation or things like that. But, for now, I would say the privacy concerns are justified.