Research suggests thinking about God makes people more likely to trust AI recommendations

People trust humans more than artificial intelligence, but when they think about God they are more likely to embrace AI recommendations over those from their peers. That's according to new research from Keisha Cutright, a marketing professor at Duke University's Fuqua School of Business, published in the Proceedings of the National Academy of Sciences (PNAS) journal.

Across eight experiments, Cutright and Mustafa Karataş of Nazarbayev University, found that "actively thinking about God" promotes acceptance of AI recommendations in a variety of contexts, including in the choice of movies, financial products, dental treatments and romantic partners.

"Thinking about God affects how we see things in our environment and our decision-making," Cutright said. "One day it hit me that AI and technology are similar in a lot of ways to religion, in that both of them are often seen as diminishing the role of humans. We wanted to see if the salience of God affects how people rely on AI."

In their experiments, the researchers randomly exposed half of the participants to tasks or experiences meant to make them think about God. In one study, they asked them to write what God meant to them. In other experiments, they used more subtle religious cues, like exposing people in the waiting room of a dental clinic to music evocative of God versus secular music; or comparing how proximity to a place of worship influenced choices.

"And of course, we also measured the extent to which participants were religious or not in the first place," Cutright said. "Because if you are religious, it's more likely that God is salient to you."

The results showed when people are actively thinking about God, they are less averse to AI and more willing to consider AI-based recommendations.

"Regardless of whether or not we're religious, we know that God is associated with a sense of power, vastness and wonder," Cutright said. "This awe for the divine elicits a sense of smallness and fallibility in people."

Cutright said the findings don't completely reverse people's overall preference for human recommendations, but they certainly show that when God is salient, humans don't have a big advantage over AI.

"The preference for the human is pretty strong, but it starts to dampen a bit when you evoke thoughts of God," she said. "It's still not going to flip to the point where people have a preference for AI over humans, but now it is probably closer to 50/50 in many situations."

She also said the studies found that in general, religious people prefer AI recommendations more than non-religious people do. Cutright said companies should consider these findings when they target certain markets.

"Counter to most people's intuition, areas with more religious populations may be good places to start with AI-based recommendations," she said.

She also said businesses using AI should carefully craft their communications.

"They could use subtle reminders of the smallness of mankind, whether through cues associated with God directly, or other awe-inducing stimuli," Cutright said. "This may enhance consumers' openness to AI-based recommendations."