Adoption and Application Trends for Generative AI at Work

In this episode:

It's been just over a year since ChatGPT first captivated the public's attention pushing many to consider the possibilities and implications of generative AI. There's been a lot of buzz about how emerging AI tools will revolutionize the working world, but how much has actually changed? What are adoption rates looking like so far? And what can practitioners expect moving into the future?

For this discussion hosted by Bryan Benjamin, Executive Director of The Ivey Academy, we’re joined by Fredrik Odegaard, Associate Professor of Management Science at Ivey Business School. Listen in as we gauge the impact ChatGPT has had on organizations, explore what changes need to be made to create a culture that adopts new AI technologies, and project what the future of AI at work will look like.

Other ways to listen:

• Apple Podcasts
• Spotify Podcasts
• Google Podcasts

Additional Resources:

What is generative AI? (McKinsey)

Implementing generative AI with speed and safety (McKinsey)

In AI We Trust — Too Much? (MITSloan Management)

Podcast Transcript:

FREDRIK ODEGAARD: Where do the rules, and the regulations, and the norms, and the accountability, and the ethics behind it — who's going to be held accountable when it's AI-generated work?

SEAN ACKLIN GRANT: Welcome to Leadership in Practice, your source for new research, insights, and advice on critical issues in business presented by the Ivey Academy. It's been just over a year since ChatGPT first captivated the public's attention pushing many to consider the possibilities and implications of generative AI. There's been a lot of buzz about how emerging AI tools will revolutionize the working world.

But how much has changed? What are adoption rates looking like so far? And what can practitioners expect moving into the future? In this episode, we are joined by Frederick Odegaard, associate professor of Management Science at Ivey Business School. Listen in as we gauge the impact ChatGPT has had on organizations, changes in the way we work, and what we can expect from the future of AI.

BRYAN BENJAMIN: Frederick, it's great to be back with you here today. I can't believe that it's been almost a year since we last sat down to have a conversation about AI and ChatGPT in particular. So it's been a little over a year since ChatGPT really became mainstream. I think it was November of 2022. It was almost overnight a household name.

So I'm really interested to hear from your perspective how things have changed, how things evolved, and we'll spend a little bit of time maybe even taking a peak at where things are going to be going next. So let's start with where are we, and what have we learned over this past year?

FREDRIK ODEGAARD: Yeah, yeah. No. Thanks for having me back, and it's great to see you again. Indeed, it's been an exciting year. So where we are, I would say we're almost more or less where we were a year ago. I think you're right. It took us by a big storm or took mainstream by a big storm, and it had all this excitement and potential.

And there was US Senator hearings and all sorts of both excitement and anxiety about it. But I think if we look now, especially if we look on the business side and commerce side, it's not like unemployment has skyrocketed. It's not like productivity has skyrocketed. I think we're still more or less in the same space that we were a year ago.

New versions of ChatGPT, for instance, have come out. And more companies have released their own versions of ChatGPT like engines. We're still waiting for this major, major explosion in the business sense, in the business world is what I'm referring to.

BRYAN BENJAMIN: Yeah, no, it's fair enough. It's actually, I was scratching my head. And as you're talking, I was like yes, yes, yes, yes. So we'll see what happens in 2025 when we sit down again. So over the past year, while the business landscape maybe hasn't changed tremendously, what have we learned? I've been to so many events and conferences, and I don't think I went to one where AI and ChatGPT in particular did not come up, either on the main stage or in discussions. So certainly, we've learned something.

FREDRIK ODEGAARD: That's right. And I think one of the things we've learned, we still see the potential, and I think all industries are aware of that. AI will transform either the entire industry or the entire business or definitely certain tasks within them. But I think what we've learned is that it's going to take a lot longer as we do that.

So right now I would say most of the applications we see I put them more as on the education or entertainment side of things. So yes, I interact with business leaders who use AI, in particular ChatGPT to write emails, to write reports, to get drafted on computer code, to create social media content. But it's not that it has, in some sense, revolutionized their own business and so on. It has on the margins has facilitated things.

So one of the things I think we have learned is that it's going to take a lot more effort than what we initially thought. And one of the things for instance, that we just recently a few days ago was when Microsoft now has released Copilot, their Bing version of ChatGPT. It's going to be a Copilot button on the keyboard.

And I think this is one of the things that people mostly are waiting for, but oh, it should just be a button for it to do that. And right now there's still too many manual steps for us to harvest or leverage the AI technology. So we'll see how-- that promise, if that will then further boost the usage and the applications of these AI tools.

BRYAN BENJAMIN: Yeah, great points. Awareness. People learning. Trying to get a little bit more comfortable, but it definitely sounds like it's still incremental versus something that's been leapfrogging and changing things in leaps and bounds.

FREDRIK ODEGAARD: Exactly. And I think that is a natural way. It's easy when we read history books about industrial revolutions. You'll read about them in one paragraph, and that takes 30 seconds to read. And you'll think that, oh, these industrial revolutions went instantly. But if we just go back and look at how electricity was rolled out, or how the internet or the web browser, it actually took decades.

So you can go back to the late 1800s when mainstream electricity was able to be rolled out. It's still took like over 50 years before like 2/3 of the US households had access to it. And again, the applications of electricity weren't immediate. Internet sort of a mainstream web browser was developed in '89. Still took over 20 years, two decades, before sort of a majority of US adults were comfortable using the internet. And so I think it's going to just take much longer. It's going to take on the order of maybe a decade or two decades before we see true revolution in business and so on.

BRYAN BENJAMIN: A very important point. We can learn a lot from how things have rolled out in the past. So you hit on a couple of, what I call maybe time saving benefits. So writing an email or drafting interview questions. And I know leaders have been trying and testing and have had some successes with ChatGPT on some of those tasks. So can you comment on what some of the successes you've seen with regards to ChatGPT over the past year?

FREDRIK ODEGAARD: Yeah, so I mean, the main successes is that it saves you quite some time. And I do it the same. I actually don't use it for writing reference letters for my students but writing code when I first do my research, or when I first come up with research questions that I'm contemplating. I'll use ChatGPT as a intelligent or dumb colleague and just ask, hey, what about this idea? What do you think of about this?

And the same thing-- the business leaders have interacted with is they can use it to save up some of their time. And one of my favorite stories I was talking to an entrepreneur. He has this fiscal rehabilitation tool. And he was using to write blogs and social media as a marketing campaign. And he was outsourcing that.

And then what he decided to do was let's see if I just feed this content to ChatGPT and let ChatGPT generate these blogs and so on. And now he vets it for accuracy and so on. And that saved him both on the orders of a couple of thousand dollars a week. He's a small entrepreneur. And now can generate that much quicker than having to audit this social creator content and so on. The same thing on letters, writers, and so on. So you save some time basically.

BRYAN BENJAMIN: Any pitfalls you've seen emerge?

FREDRIK ODEGAARD: I think the main pitfalls or the main worries, especially these workshops that we have hosted is when we get participants from the legal and tax and medical profession. The big concern still is on safety of data or security of data. And we're still in the world where ChatGPT sort of dominates as the de facto Google in this space. But people in most professions aren't willing to actually feed in all their data on ChatGPT for the data safety reasons or security reasons.

So most companies have actually in these professions have said, no, you're not allowed to use ChatGPT because we're too concerned about the data. I think there the issue is going to be that right now it takes a lot of computing power to build your own engines. So, of course, the big companies like Amazon and Google and IBM, they can generate their own of these AI tools like ChatGPT.

But if you want to have one in-house, if you're a law firm or a tax firm, you have to, of course, make sure that you can generate these responses based on your own intelligence on all the documents that you have. And so how do you now do that? And I think that's one of the next steps will be, how can we make it sort of off the shelf that a company can have their own version and not have to be concerned about, oh, if I feed this into ChatGPT, what happens with that data?

Data security has always been an issue since the internet, and I think this is the one main pitfall still so far, I would say.

BRYAN BENJAMIN: Yeah, I've heard a lot about that, especially from some of the professions that you talked about. So I imagine we're going to see a lot more in this space in terms of research and testing and new approaches.

FREDRIK ODEGAARD: Yeah. And I think the second pitfall or concern is, I think, we still haven't figured out the accountability and the responsibility. And that's the other one. And this is one of the main themes in these workshops that we hold is as you think about having AI be part of your operations or your business strategy and so on, where do the rules and the regulations and the norms and the accountability and the ethics behind it, who's going to be held accountable when it's AI-generated work?

The ethics of AI has been a big concern or topic within the media and so on. And here it seems like everybody's just punting the decision on the other ones. The industry makers like the OpenAI companies and the Elon Musks of the world they are pushing it on to, oh, the government and the regulators have to come up and let us know.

And they on the other hand, well, we have no idea. And they're pushing it back on the technology people. Tell us what we should do, and how should we frame the boundaries of the regulation and so on. And that's one thing that we're still not quite clear of what is it we should do.

And I think-- I mean, if I were to-- I think one thing that I see who will be first-- again, I'm just making an educated guess here. I think maybe the first place we'll see it is in the tax code. I think maybe companies like the Canadian Revenue Agency or the IRS, the American version of it, I think because they are so strict on rules and boundaries, so it wouldn't surprise me that the initial movement in trying to establish boundaries with where does the AI accountability layer and so on will be on tax codes and tax declarations.

BRYAN BENJAMIN: Interesting. Could you just take a few minutes and walk us through how you think about them.

FREDRIK ODEGAARD: Yeah. Yeah, no, this is for sure one of the-- I think, this is one of the biggest confusions out there. And I agree. Because I think different people have different ideas of what is it that's AI. What is it that's ChatGPT. What's analytics. What's machine learning and so on. And I think you basically have to just whatever industry or whatever company business you're working in and you have your own vocabulary around it.

So I can't really give an answer to that. But I think one of the things that helps me-- what I think makes it is, are you focusing on the technology, on the servers, and the wires, and the software and the hardware, or are you focusing on the abilities? If it's analytics or marketing or operations or finance, what are you specifically focusing on when you talk about AI and so on and ChatGPT?

So I mean, ChatGPT is right now just this interface that we see, and it can generate human-like responses. But you're right. In some sense, AI is perhaps even broader and includes the cloud computing and all the wires and the sensors that are self-driving cars. So yeah, it's not quite clear actually.

[MUSIC PLAYING]

BRYAN BENJAMIN: We're going to talk about leaders now. And what do leaders need to learn in order to support their teams? As you say the word "embrace," it's coming. So leaders need to ensure their teams are comfortable and ready to be able to take advantage of the opportunities that come with AI and tools like ChatGPT. What can they do?

FREDRIK ODEGAARD: Yeah. So you're right. And I think we talked about this a year ago. One thing you cannot do is fight it. Forget about that. And I think we discussed this a year ago too, but you certainly at some level have to embrace or have to accept that this is coming and so on.

And with that, you're probably going to even have to accept that there is going to be a setup cost. There's going to be some initial investment, both perhaps in terms of technical infrastructure, as well as in learning and education for your workforce and changing the processes by which you do work.

And so on that's one thing but just be ready to accept that. And I think that's maybe why it's going so relatively slow. I would say is that it's easy if you think about, oh, I can have AI automate some email responses. OK, but these email responses maybe it takes me 5 minutes, and it's going to take me maybe two days to figure out a process by which I'm going to now have an automated system.

And people are so busy these days. Like you can't really think about, oh, take out two days just to solve this 5-minute problem. But yeah, but the 5-minute problem is a recurring 5-minute problem every single day or twice a day and so on. So to embrace the setup cost with it.

Second part is to actually think deeply about what we talked about the norms and the accountability and the like who is it that owns the tasks that gets done by AI in terms of accountability. And then the second part would be-- and this is in certain sectors of the business world. And again, we bring these up in our workshops is just because AI can do something, doesn't mean that you as a business organization want it to do something.

And so we discuss about where are the companies commercial interests, and where is the privacy of customers, where is the privacy of your employees? So just the fact that you can use this technology to figure out a whole bunch of things, or maybe come up with strategies.

Maybe you as a company, as a leader have to decide you know what? Yes, I can see there's perhaps I can gain an insight an advantage, but morally or ethically I don't think it's right. I think we should just not do that. And there's all sorts of privacy issues that, I think, leaders need to think about.

BRYAN BENJAMIN: And I agree with the comment around you can't hide. It's coming. It's here. It's going to be coming more and faster. So what can you do to help your teams embrace it? And you hit on some of the-- I think going a little bit further than just testing it out is where can we go next.

FREDRIK ODEGAARD: Yeah, exactly set up like an automated process so that button that Microsoft is now introducing actually works.

BRYAN BENJAMIN: So speaking of supporting leaders, and you've referenced this a little earlier on, is we've been running a workshop that you designed and have been facilitating as our faculty lead that is focused on helping leaders functionally understand how to look for opportunities to integrate ChatGPT into their daily work tasks and support teams to try this as well.

So we've hit on a little bit of this as well. But I'd like to hear from you in terms of what you've learned from the experience in working with those participants in the workshop.

FREDRIK ODEGAARD: Yeah, no, it's been absolutely phenomenal. And I think one of the things-- because we draw from such a wide spectrum of industry. We do get the tax accountants. We get the lawyers. We get from medical professions. We get from marketing. So that's just been so fascinating to hear their take on one, how they can, after the workshop where they see the opportunities that they now can use and just the insights they have gained.

We had someone from local government involved in one of the workshops. And just again, from a public office perspective, think about what are going to be the implications of AI. So I think that's been the most rewarding for me is to see wow, the use cases of how they perceive, oh, this is where AI is going to take on and happen and stuff like that. So that's been very, very fascinating to see actually.

BRYAN BENJAMIN: In different industries, different types of organizations, different geographies, any themes in terms of why people signed up for the workshop to participate? Like, what are you hearing?

FREDRIK ODEGAARD: I mean, so besides the buzz, of course. There was just this we certainly hit a good tune there because it was so contemporary. It was everywhere. And so everybody had heard the buzz and just wanted to know more about it. But beyond that, I think-- and I think this is basically Ivey Academy is they just wanted to grow further as leaders.

And you just know that I need more tool sets than just-- if I'm a marketing person, I need more. I can't just focus on my marketing expertise. If I'm on operations, I can't just focus on-- I need to be much more well-rounded. And here's something that's clearly going to revolutionize all industries. And so I just need to broaden my skill set. I think that's been one of the main general takeaways.

BRYAN BENJAMIN: With that consistent lifelong learning, theme is always need to be kind of updating and upskilling and seeking what's new and what's next.

FREDRIK ODEGAARD: Yeah.

BRYAN BENJAMIN: So same thing. In the workshops, any sort of themes emerge around challenges that leaders are facing?

FREDRIK ODEGAARD: Yeah, so I think the big one is the one we've talked about, the data security, especially when we have the legal, the tax, and the medical professions. They are very, very concerned about how do we make sure that our clients records are kept. How we make sure that medical patients file are not breached and so on. That's been the main thing there.

And then it's just about trying to think outside the box. And, I think, every time new technology comes along, including electricity, including the internet. When it first comes, we're like, all right, what is this? And what do I do with it? And we can only play around on the fringes.

But then after just playing around a little bit, we make these big jumps, oh, here's an opportunity. And we just see the world open up much wider. And I think so we're still trying to explore in all these fields, different industries. It's like, OK, what's going to be order of magnitude improvement from what we're doing? That we're still waiting for actually to see the revolution itself.

BRYAN BENJAMIN: So we started reflecting on-- when we sat down at the beginning of last year. Let's flip that and say we're sitting down at the beginning of next year. What do you hope we're able to talk about? What's happened in those 12 months?

FREDRIK ODEGAARD: That's right. Yeah, that will be fascinating. Yeah, we should have asked that question a year ago. Like, where are we going to be a year from now? Yeah, I don't know. I'm hoping at least that we will have some of the legal regulations figured out and particularly about the accountability as I point out.

I hope companies have maybe been able to come up with some sort of code of conduct of AI and where they want to draw their boundaries. In terms of technology, I think we already see-- so as I talked about, so OpenAI the company has their own engine. IBM, Amazon, Google, and Microsoft.

And then a lot of the big consulting firms like the big ones they have their own engines. I think we're going to start seeing off the shelf that you can just go to store or just download in fact. Oh, let me just download my own version, and I'm going to build my own large language model and build my own version. So that might the technology side see it.

BRYAN BENJAMIN: All right, well, we now have it. So we can check 12 months from now how close we were. So in terms of order of magnitude, are there any industries in particular that you feel maybe hit harder or faster or see new opportunities emerge quicker than others?

FREDRIK ODEGAARD: Yeah. So I don't know if I can pinpoint a particular industry itself. But I think one of the paradoxes that we will see is that we tend to think about AI and analytics and to be on all about the number crunching and the ones that are already super technologically advanced. But I think the big jumps are actually going to come from the more what I would call the qualitative work.

And so the more qualitative your work is and the tasks that you perform, that is where we're going to see the big jumps first. I think the ones that are already very technologically sophisticated, such as the finance industry perhaps or logistics and so on, I think that's where we will see incremental improvements or applications of AI in the beginning.

But the big ones are going to come in maybe counseling, for instance, something in the social welfare space. But the more qualitative the tasks are and the industry, that's where we will see the true revolutions first and the big, big jumps. And that's sort of a paradox. That's not what most people think about.

BRYAN BENJAMIN: That's an interesting perspective. I hadn't given it that kind of thought. But you're right is you sort of assume it's ones where there's more tasks that can be automated, but these actually are areas where it could be a little bit more subjective or complex that could see the biggest opportunities, benefits changes.

FREDRIK ODEGAARD: And again, it's hard for us to see because it's hard for most people to think outside the box of what they already do. But that is truly where I think we will see the big leaps immediately with using AI technologies, the more qualitative, the more manual aspects of human or commercial activities.

BRYAN BENJAMIN: You think about the potential implications. So I'll use your example to go off of qualitative role, a counselor. A number of organizations provide access to counseling through benefit providers. Is my counselor a person, or is my counselor a computer, or is it a person aided with a computer, and how does that change the game?

FREDRIK ODEGAARD: That's right. And I think that is where we will see the biggest impact or benefit is when it's a person supported by AI as a counselor, for instance, and able to give more personalized recommendations on whoever they're dealing with.

BRYAN BENJAMIN: And that's a theme I've heard you comment on throughout, which is the it's not this dichotomous it's either this or it's that. It's can it enable and support actually work in tandem with someone.

FREDRIK ODEGAARD: That's right. Business and commerce is part of humanity. It's part of society. And so you can't outso-- you can't automate out the human value or the value created for humans in a business organization. So if you automate a task, both on the production of the service and on the consumption of the service, what value does that add to a company?

We always have to focus on that this is going to help us as humans to make our lives better and more efficient and so on the commercial side. There's a funny joke. One of these interview joke-- I interviewed singers or guitar players in The Eagles. And it's all about the revolution of AI in music production and so on. And so we asked him, so what's your thoughts about AI in the music world?

And he chuckled, and he said, ah, I'm not so concerned. An algorithm can never destroy a hotel room. So that's funny, of course, right? And so you're thinking about the sex, drugs, and rock and roll lifestyle. But I think there is actually something to that. And that is that the algorithm will never occupy a seat down to a vacation in Puerto Vallarta. An AI will never consume a sandwich made and so on.

So yes, there are some tasks that we will automate. But fundamentally, we have business as part of society. When we talk about value creation, it is for us. And so I think this is where we will see that-- I don't think we can-- we can't have AI destroy value in terms of humanities.

SEAN ACKLIN GRANT: Thank you for tuning in to Leadership in Practice. We'd like to thank our guest Frederick Odegaard. Leadership in Practice is produced by Joanna Shepherd and me, Sean Acklin Grant. Editing and audio mix by Carol Eugene Park.

If you like this episode, make sure to subscribe. You can also find more information by visiting iveyacademy.com, or follow us on social media at IveyAcademy for more content, upcoming events, and programs. We hope you'll join us again soon.

[MUSIC PLAYING]