Chatbot or human? Either way, what matters for customer trust is “perceived humanness”

The helpful person guiding you through your online purchase might not be a person at all.

As artificial intelligence and natural language processing advance, we often don’t know if we are talking to a person or an AI-powered chatbot, says Tom Kelleher, Ph.D., an advertising professor in the University of Florida’s College of Journalism and Communications. What matters more than who (or what) is on the other side of the chat, Kelleher has found, is the perceived humanness of the interaction.

With text-based bots becoming ubiquitous and AI-powered voice systems emerging, consumers of everything from shoes to insurance may find themselves talking to non-humans. Companies will have to decide when bots are appropriate and effective and when they’re not. This led Kelleher — along with colleagues at UF, California Polytechnic and the University of Connecticut — to develop a measurement for perceived humanness. They shared their results in the journal Computers in Human Behavior.

In the study, participants chatted with bots or human agents from companies like Express, Amazon and Best Buy and rated them on humanness. Sixty-three of 172 participants couldn’t identify whether they were interacting with a human or a machine. But whether the interaction featured AI or not, higher scores of perceived humanness led to greater consumer trust in the companies.

“If people felt like if it was human — either with really good AI or with a real person — then they felt like the organization was investing in the relationship. They’ll say, ‘Okay, this company is actually trying. They’ve put some time or resources into this, and therefore I trust the organization,’” Kelleher said. 

Kelleher started studying how language affects customer trust more than a decade ago, when blogging culture introduced a conversational approach to the stuffy, stilted language businesses tended to bludgeon their customers with. Companies noticed that as jargon waned, consumer trust, satisfaction and commitment grew. The new study shows that the same holds true with chatbots and other online interactions, and can be applied to bots and humans alike. (“An agent can be so scripted that people feel like they’re talking to a machine,” he explained.)

As AI-powered interfaces blossom, even expanding to include animated avatars that look human, ethical issues will follow. Should companies disclose when customers are interacting with a non-human agent? What if the helper is a hybrid: a person assisted by AI? Are there areas where consumers won’t accept bots, such as health care, or situations where they might prefer a non-human?

“If I’m just trying to get an insurance quote, I would almost rather put something into an app then have to make small talk about the weather. But later on, if my house floods, I’m going to want to talk to a real person,” Kelleher said. “As the metaverse evolves, understanding when to employ AI and when to employ real people will be an increasingly important business decision.”