New Research: You Think They Used AI, They Think You Did

The AI Trust Gap: What New Research Reveals

That nagging suspicion that everyone else leans on AI more than you do isn’t just in your head—it’s measurable. Zoe Purcell, a cognitive scientist, led two online studies with over a thousand adults to find out what people actually believe about AI-assisted communication.

Participants watched realistic scenarios staged inside familiar apps like Gmail, spanning text, audio, and video, and ranging from no AI use to subtle AI use to heavy AI enhancement. Then they answered two questions: how likely are you to use AI in a similar interaction, and how likely is everyone else? On a 100-point scale, participants put other people at roughly 71—but themselves at roughly 40.

Purcell’s team also found that openly disclosed AI use was considered more acceptable than hidden use. In other words, the secrecy bothers us more than the tool itself. We hold others to a harsher standard than we apply to ourselves, and we assume the worst about what happens behind the scenes.

Why Defaulting to Distrust Is an Emotional Intelligence Problem

Emotional intelligence (EI) is your ability to recognize and understand emotions in yourself and others, and to use that awareness to manage your behavior and your relationships. It breaks down into four core skills: self-awareness, self-management, social awareness, and relationship management. To see how you score, experts recommend taking a self-assessment that assesses each of the four core skills.

Defaulting to distrust is problematic. It signals a default to distrust and doubt, when the research shows that defaulting to trust builds faster and stronger relationships. In The Speed of Trust, Stephen M. R. Covey argues that trust isn’t a soft virtue but rather an economic one. When trust is high, speed goes up, and cost goes down. When trust is low, everything slows down.

Asymmetric Feedback: The Psychological Reason We Assume the Worst

Why do we default to distrust when people are, in fact, more trustworthy than we think? Psychologists Detlef Fetchenhauer and David Dunning ran an experiment in which participants had to decide whether to trust 56 strangers in an economic game. Participants predicted that only 52% of those strangers would prove trustworthy and return the money they were given. In reality, 80% did.

The culprit behind this drastic gap is what’s called asymmetric feedback. If you trust someone and get burned, you receive immediate feedback that you were wrong to trust. But if you withhold trust, the person never has a chance to come through, and you never get to confirm whether you made the right call. The evidence only ever piles up on one side.

Over time, this asymmetric feedback encourages you to default to distrust. You rarely learn about the trust you could have extended, so suspicion starts to feel like the safer, more rational choice—even when the people around you are far more trustworthy than you assume.

The EI Strategy: How to Default to Trust in Three Steps

To counter your default toward distrust, use these three steps next time you feel suspicious of someone.

1. Catch yourself at the moment when you begin to get suspicious. Suspicion is a thought, but drawing a conclusion based on that thought is an active decision. Try to start noticing those moments where you assume something about someone (not just with AI use!).
2. Apply your own standards to others. You know exactly why you would or wouldn’t use AI. Ask what you would want assumed about you if you had sent that exact message.
3. Write a more generous version of the story in your head. Put together one plausible, charitable explanation for the other person’s behavior and take it seriously.

This concept extends far beyond playing the role of AI detective. It’s about noticing when you default to distrust and then beginning to redirect your thoughts toward a place of trust and curiosity.