By Jamie Merchant

The hype has been extraordinary.

Over the past two years, news outlets have blanketed the public with stories about the impact of large-language models (LLMs), or “artificial intelligence,” and their profound implications for human civilization. The CEO of Tesla and billionaire investor, Elon Musk, warns that the technology represents one of “humanity’s biggest threats.” Other commentators predict a more benign future in which AI liberates us from toil, taking over the mundane tasks of office work.

With conflicting reports like these, one could be forgiven for feeling confused.

Putting the hype aside, what are the facts on the ground? How is this emerging technology actually used by businesses and organizations?

At DePaul’s Driehaus College of Business, new faculty in the Department of Marketing are cutting through the headlines to investigate the promise, and the limits, of LLMs for modern businesses.

One innovative use of LLM’s is for producing “synthetic data”, AI-generated responses that simulate humans in order to inform business intelligence. “How good is AI at really representing human variety?” asked Ignacio Luri, an Assistant Professor of Marketing at Driehaus. “That’s something I’m skeptical about. But it’s happening, so I’m studying it.”

Luri, whose background is in marketing and linguistics, also studies what he calls the “market conversation”: the dialogue that unfolds around companies, consumers, and the brands that connect them. His current research focuses on the uses of AI analytics for studying that conversation. “I mostly study big data,” he said, referring to the modern study of human behavior based on very large data sets. “But I also have a qualitative toolkit.

“The market conversation is very cultural at heart,” he explained. “It happens in a cultural context. It can be really tempting to take a dataset and just crunch the numbers. But we’re talking about people. When we’re talking about the consumer conversation, things happen in a context – always. Who said that? When? Why? In what context? To whom?”

In other words: how consumers see brands, and conversely, how companies understand their customers, are both products of an ongoing dialogue between them. And, like any dialogue, the market conversation unfolds in the assumptions, habits, and beliefs that characterize particular people in a particular community at a particular point in time – that is, in all the messiness of human communication.

How apt are large language models to capture the subtle nuances of human speech, or the unique meanings that attach to specific words for a given community? Amidst all the enthusiasm for AI, Luri insists on the importance of not losing sight of the human element – the inherently contextual nature of communication.

“All that matters. It’s important not to abstract away from all that reality.”

 

As with generative AI, buzzwords swirl around the emerging field of neuromarketing. According to the Harvard Business Review, this new field “studies the brain to predict and potentially even influence consumer behavior and decision making.” One can easily imagine the value of this technology for businesses.

Such is the potential, but how does it work in practice? Most importantly, what are the real advantages and limits of the technology?

“With neuromarketing, I always say it’s a supplement, not a replacement, for traditional marketing research,” said Jennifer Tatara, Assistant Professor of Marketing at Driehaus.

Tatara works at the cutting edge of the neuromarketing field, which mines insights from psychology, marketing, and economics to look at the science behind consumer decision-making. Neuromarketing introduces biometric data into the study of consumer behavior, assisting researchers with an age-old question: what motivates people to make the decisions they do?

Marketing researchers and professionals, of course, are interested in a specific subset of people: consumers.

“We can use these tools to see into the decision-making process in a different way,” said Tatara, “to get into the black box of decision-making. We need a wide range of tools to get the full picture. But without biometric tools, you’re missing a piece of that picture.”

Some of these tools might be familiar. Electroencephalograms, for instance, gauge mental activity by tracking the small electrical impulses given off by the brain in response to stimuli. But some are more exotic: eye-tracking software yields insights into where consumers’ focus is drawn. Galvanic skin response, a measure of minuscule amounts of sweat, correlates with subjects’ emotional arousal.

Tatara emphasizes the potential of these tools to help both consumers and businesses make better, more satisfying decisions. But she is also quick to deflate the exaggerated claims sometimes made on its behalf, and to point to the ethical dimensions of this new field. As she puts it, “there’s no magical ‘buy’ button in the brain – this isn’t mind control.”

“Like with any new tool, there are ways to use it positively and ways to use it negatively. As marketers with access to these tools, it’s our job to make sure we’re not only selling, but we’re also helping. With a better understanding of how people make decisions, we can help them make better decisions.

“That’s why I’m happy that DePaul is taking an active role in teaching these tools. Here, faculty and students study and apply these tools; we’re doing it ethically, and we’re doing it to help consumers, at the end of the day.”