Wharton@Work

February 2023 | Management

Don’t Fear AI — Collaborate with It

Since it was released late last year, ChatGPT has been described as portending the end of journalism, business consultancies, and teaching. Wharton@Work spoke recently with marketing professor Raghu Iyengar about the fear evoked by the AI chatbot and the possibility of organizations using it, and other AI technologies, to enhance their business. The director of the Analytics for Strategic Growth: AI, Smart Data, and Customer Insights program is decidedly positive.

W@W: So much of the reaction to ChatGPT is fear-based. Do you think the fear is warranted?

RI: I think we are witnessing some very interesting times. There are a number of new generative AI tools, including ChatGPT, that produce content that is not original in some sense. These tools present a massive opportunity, but their use is not without risk. That’s what CNET, the tech news site, found when they used AI to write dozens of articles. While the articles made a lot of sense, they also included factual errors, which they've had to correct or retract. Additionally, it now seems that the AI tool has also plagiarized content from past articles on CNET and its competitors. On the one hand, generative AI can offer an opportunity: for instance, if a part of your day-to-day activities is writing articles that summarize past content, it can likely be taken over by what I call an augmented AI companion. But you also get more “reflection time” to engage in activities that can be helpful in developing your career.

W@W: How should leaders consider the opportunities versus the risks?

RI: As a decision maker, you need to step back and look at the kinds of things you do in terms of either interacting with customers externally or interacting with your own employees internally. Of those activities, which are the ones that perhaps can be better done by an augmented tool, helping free up time for people in the organization to become more creative, getting involved in activities that may be very hard for an AI tool to do? I call them “reflection activities,” which involve making sense of what's going on out there.

An interesting analogy is the emergence of ATMs. When they came out in the marketplace, everyone was nervous that the job of the bank teller would go away. That's not what happened. If anything, they freed up bank tellers to be involved in interactions with customers that go beyond just giving them money or depositing their checks. They can be involved in more personalized, engaging activities and build relationships with customers.

W@W: What are some other kinds of opportunities that people might not be thinking of or can’t see because they are wary of AI?

RI: Before the pandemic there was a constant call for people to retool themselves in response to the number of opportunities in the digital landscape. It was a recognition of the need for constant learning. Consider the role of a CMO, for example. It has evolved to include engaging customers, managing the digital pipeline, and making all of the data that’s being collected more actionable. It has become more integral to organizational strategy, and a lot of functional boundaries are becoming blurry. Marketers have to become more familiar with more aspects of the business. And so retooling, getting new education, and keeping up with the times is important. Many of the new AI tools are also nudging us in that direction, because if we are starting to use some of these tools for taking on day-to-day activities, then we need to be equipped to go beyond that.

Having a growth mindset, a desire to constantly keep learning, is critical. People might consider if part of their responsibilities can be handled by a tool, what are the other parts they can lean into? How can they become more creative, taking on more responsibilities that perhaps showcase the human element in the human-AI collaboration?

W@W: That collaboration and need to get out of your comfort zone to make it happen is really interesting. It’s not about AI tools simply automating certain functions but about using it in combination with human insights and expertise to drive business outcomes. And it sounds like AI can, in that way, help make the human job more interesting, creative, and fulfilling.

RI: Indeed. There can be a happy medium between automating everything, which many companies have done, and being completely human-centric. This could mean having AI generate a bunch of alternatives for your customers, and then whittle them down to a smaller number. But the human in the loop can then decide the particular fit between the fewer alternatives and the customer. I think it's this human-AI collaboration that could lead to a win-win as opposed to being purely human-centric or purely AI-centric.

W@W: Can you say more about the need for a learner or growth mindset? It seems to be a real key to shaping the potential threat of AI tools into opportunities.

RI: Selling a business idea, whether it is about analytics, marketing, or finance, is a combination of analysis and being able to tell a story around it. So, being able to justify how the latest AI tools could integrate with the company's objectives, for example, requires not just analysis, but also the ability to explain how people will use it to make better decisions.

That is why developing your skills as a storyteller or data translator has always been part of the Analytics for Strategic Growth: AI, Smart Data, and Customer Insights program. It's not a question of understanding or explaining how ChatGPT works. You need to know what ChatGPT can do for your organization that was either not being done or was being done by someone else. And if that someone else is now going to be given additional responsibilities when an AI tool takes over some of their previous ones, how can we better equip that person?

Part of this story needs to be told and understood from a human resource perspective, asking how you can make sure that employees actually do have higher personal and job satisfaction when some of their responsibilities become automated. What kind of education should you offer, and how can you give them some time to consider what else they can do to contribute?

W@W: What if you can’t make a business case for AI tools right now?

RI: One has to think more broadly about strategic concerns and always tie any decision with business objectives. As an example, consider the case of dashboards, which have been with us for a while. As more data is collected, there are more dashboards, but then how do we decide which ones to focus on? As new tools come along, having the sense of the relationship between the tool and the business objective becomes even more important.

We also need to keep in mind that some of those tools may be a great fit today, and some may be a better fit later on, depending on the maturity of the organization. Beware of what I call Shiny Object Syndrome. Today it's ChatGPT, and tomorrow it might be something else. When you are looking at the latest tool, you need to consider how it can integrate into your business, and what it can do for you that furthers or complements what you are already doing. Can it be used as a collaborative tool? Once you think of it as a complement, then it becomes perhaps an opportunity; in that way, these tools can potentially help you achieve even more.