Serving Strangers in a Virtual Restaurant

Virtual reality has been sold as an “empathy machine” for more than a decade: Slip on a headset, inhabit someone else’s body, and (or so the promise goes) your biases will melt away. Reality, however, has been complicating that story. A recent study on intergroup contact in virtual reality suggests that VR does not dissolve prejudice so much as rearrange it, sometimes in unexpected ways.

The study placed participants inside a virtual restaurant. They were servers, embodied as avatars, interacting with a virtual customer. Sometimes the customer was friendly, sometimes rude. Sometimes the customer appeared to belong to the participant’s racial ingroup, sometimes not. And sometimes the avatar in the mirror looked exactly like the participant, while other times it bore a stranger’s face.

On paper, the setup sounds simple, but in practice it turned VR into a laboratory for testing a subtle question: When does contact soften group boundaries, and when does it reinforce them?

Embodying Yourself (or Someone Else)

One of the most intuitive ideas in VR design is that realism breeds responsibility. If your avatar looks like you, you may behave more thoughtfully, knowing that others could recognize you. This is called perceived identifiability, the feeling that your online self is traceable to your offline one.

The study found that this intuition is only partly right: using a self-resembling avatar did not automatically reduce prejudice. The decisive factor was whether participants felt identifiable. When they did, the boundary between self and other began to blur, and participants reported greater “self-other overlap” with the virtual person they interacted with.

This effect was strongest among people who had higher social dominance orientation, a personality trait associated with preference for hierarchies. In other words, the very individuals often assumed to be resistant to prejudice-reduction efforts showed measurable psychological shifts when identifiability was high. That finding challenges a common assumption in diversity interventions: that those most invested in hierarchy are least likely to change. VR suggests a different possibility: under the right conditions, accountability (not anonymity) may pull even resistant users toward psychological closeness.

The Surprising Power of a Stranger’s Face

Yet the study also delivered a twist: When participants interacted with an outgroup member while embodying an avatar with a stranger’s face, they showed increased acceptance of the outgroup afterward.

Why would anonymity help?

One explanation comes from classic social psychology: Distancing oneself from one’s ingroup identity can lower anxiety. A stranger-faced avatar may dampen the instinct to defend one’s group status, creating a more neutral psychological space. Rather than feeling like “me versus them,” the interaction becomes “a person doing a job with another person.”

This matters because it suggests that VR designers face a trade-off: While personalization can increase accountability, depersonalization can reduce defensiveness. The effectiveness of VR contact may depend less on realism per se and more on how identity is strategically dialed up or down.

When Bad Encounters Do Good Things

If positive contact reliably reduces prejudice, negative contact should reliably increase it—or so the standard theory goes. In this study, negative interactions with outgroup members unexpectedly increased empathy toward that group. The participants, embodied as service workers, may have internalized the norms of their role. The rude customer did not become a symbol of the outgroup; Instead, participants seemed to interpret the behavior through situational lenses familiar from everyday life.

This does not mean negative contact is harmless, but it does mean that context matters. VR does not exclusively replay social encounters; it assigns users roles, scripts expectations, and shapes moral interpretations. A single rude interaction can sometimes provoke understanding rather than backlash, especially when the user’s role invites patience rather than power.

Bias Essential Reads

Your Child Isn't the Problem. Their School Report Might Be.

Is Mandated Reporting Racist? What Families Must Know

The larger lesson from this work is not that VR “works” or “fails” at reducing prejudice, but that VR is exquisitely sensitive to design.

Small choices (whose face you wear, whether you feel recognizable, whether the interaction is courteous or tense) cascade into different psychological pathways. VR contact can foster closeness, defensiveness, empathy, or retrenchment, sometimes all at once.

If virtual reality is to be used responsibly in education, training, or social intervention, it must be treated less like a moral shortcut and more like a social instrument to be tuned, misplayed, or composed. The headset does not erase prejudice, but under the right conditions, it can reveal how surprisingly malleable (and stubborn) our group boundaries really are.