When “I don’t know you yet” already feels like a warning
A stranger’s avatar walks up to you in VRChat and says hello.
One person starts chatting. Another thinks, “Who are you, and why are you talking to me?”
Same platform. Same greeting. Completely different experience.
The key difference is often not social skill. It is what “unknown person” means at the starting line.
For one person:
trust = 0 means “not evaluated yet.”
For another:
trust = 0 means “stay alert.”
That changes everything.
The “mask” metaphor is useful here, but not because avatars literally look like masks. The mask is psychological: you cannot see the person’s motives, history, real identity, or reason for approaching you.
1. Trust always contains some uncertainty
Trust means accepting a little vulnerability while you do not fully know what the other person will do.
Even a casual conversation with a stranger contains uncertainty. They may be friendly. They may be annoying. They may push boundaries. They may be recording. They may simply say hello and leave.
The hard part is that you usually do not know the probabilities.
A 2018 Nature Communications study found that people who tolerated ambiguous uncertainty better were more willing to trust and cooperate with strangers. Once information about the other person’s behavior became available, that individual difference became less important.[1]
So “I don’t trust strangers” is not automatically irrational fear. People differ in how comfortable they are acting before the other person is predictable.
2. Social VR creates an unusual combination: high presence, low background information
VRChat can feel socially close. Someone has a body, a voice, a direction of gaze, and can stand right in front of you.
But you may know almost nothing about them.
On a conventional social network, you can often inspect a profile, posts, mutual friends, or history. In a workplace or school, you at least know why the person is there.
In a public VRChat instance, a colorful avatar can simply walk up and speak.
A 2024 study comparing social VR with other communication channels found that social VR offers strong interpersonal affordances while also being characterized by notable privacy concerns and reluctance to trust other users.[2]
In other words: the person can feel close before they feel known.
3. It can feel like an unknown international phone call
A useful analogy is an unexpected call from an unfamiliar foreign number.
Who is this? Why are they calling? How did they get my number? Is this a scam? What do I gain by answering?
Many people do not think, “I’ll answer first and find out whether it is safe.”
They think, “Because I do not know whether it is safe, I will not answer.”
For some VRChat users, an unsolicited greeting from a stranger is psychologically closer to that than to meeting a friend of a friend at a party.
That creates a loop:
“I need information before I talk.” But: “I would need to talk to get the information.”
A perfect chicken-and-egg problem.
4. Two different ways to build trust
One strategy is certainty-first:
unknown person → unclear intentions → insufficient trust → avoid interaction
Conversation happens after the person already seems reasonably safe.
The other strategy is behavior-first:
unknown person → low-stakes conversation → observe behavior → continue if it feels good → leave if it does not
Here, conversation itself is reconnaissance.
Research on online acquaintance formation found that direct interaction was especially effective at reducing uncertainty about a new person.[3]
That explains why one person naturally says, “If you don’t know them, just talk to them.”
And another naturally answers, “If I don’t know them, why would I talk to them?”
They are solving uncertainty at different stages.
5. “I can just block them” is reassuring only if exit feels cheap
VRChat provides mute, avatar hiding, blocking, safety settings, and other user controls.[4]
For some people, that radically changes the risk calculation.
“If this gets weird, I leave.” “If they annoy me, I block them.” “If the room feels bad, I change worlds.”
The goal is not to prove that the stranger is safe before speaking. The goal is to keep the cost of being wrong small.
But this does not reassure everyone.
For another person, one unpleasant interaction is already a cost. Being approached is already a cost. Having to decide whether to block someone is already a cost.
So “you can always block them” is not a universal answer. Different people are counting different costs.
6. The same avatar can be a safe costume or a suspicious mask
Anonymity has two opposite psychological effects.
It can reduce trust:
“I cannot see who you really are.”
But it can also increase comfort:
“You cannot judge my real appearance.” “I do not have to expose my offline identity.” “I can try a different way of presenting myself.”
Research on VRChat and social VR has documented exactly this identity flexibility and, for some users, the sense of safety that anonymity can provide.[5][6]
So the same avatar can mean:
“I am hiding from you.”
or:
“I am protected enough to be myself.”
That contradiction is not a bug. It is one of the defining social features of VR.
7. Anonymity does not automatically make people dangerous
It is tempting to jump from “anonymous” to “unsafe,” but the evidence is more complicated.
A 2025 social VR experiment found that anonymized generic avatars did not produce greater disinhibition than self-similar avatars. Gaze and turn-taking still resembled face-to-face interaction.[7]
Earlier online research also suggested that “anonymity” is too broad an explanation for bad behavior; factors such as lack of eye contact can matter independently.[8]
So the most accurate starting point is not:
unknown avatar = dangerous person
but:
unknown avatar = less information, therefore more uncertainty
For people who dislike uncertainty, that alone can be enough to stop interaction.
8. Why the same pattern can appear on the phone
This is why the behavior may not be specific to VRChat.
An unfamiliar phone number creates the same basic structure:
unknown identity
- unclear purpose
- low context = high uncertainty
Research on communication with strangers has linked uncertainty and anxiety with avoidance.[9]
So someone who dislikes speaking to unknown callers may react similarly in VRChat: the entry barrier comes before the actual conversation.
9. This is not necessarily “mistrust of humanity”
If someone becomes perfectly comfortable after they know you, the issue may not be generalized distrust.
It may be distrust of uninitialized relationships.
They may need one or more of these before engagement feels worthwhile:
- a shared friend
- a familiar community
- some visible history
- time to observe first
- slower pacing
- clear control over leaving
That is different from simply being “bad at talking.”
10. The real difference is what trust zero means
The shortest model is this:
Person A: trust 0 = unknown → interact a little → update trust from behavior
Person B: trust 0 = caution → wait for evidence → interact after trust rises
Person A builds trust through conversation.
Person B wants enough trust before conversation begins.
Both strategies are internally coherent.
11. Why the “mask” metaphor works so well in VRChat
Offline, people already edit how they present themselves through clothing, titles, profiles, and social roles.
VRChat turns that dial all the way up.
You can be a cat, a robot, a fantasy character, or anything else. Voice and appearance do not have to match. Real-world identity can remain irrelevant.
For one person, that is liberating:
“I do not care who is inside the costume. If the conversation is fun, we talk.”
For another, it makes uncertainty worse:
“If the outside can be anything, I know even less about who is inside.”
The same system increases freedom for one person and uncertainty for another.
That is probably the heart of the difference.
The real question is not only, “Do you trust people?”
It is:
“How safely can you leave an unknown person in the category of ‘not evaluated yet’?”
And:
“Do you believe you can exit quickly if the interaction goes bad?”
Those two answers largely determine whether a VRChat stranger feels like someone in a costume or someone behind a mask.
References (9)
- Vives, M.-L., & FeldmanHall, O. (2018). Nature Communications, 9, 2156 doi.org
- Cummings, J. J., & Ingber, A. S. (2024). Computers in Human Behavior, 161, 108427 doi.org
- Antheunis, M. L., Valkenburg, P. M., & Peter, J. (2010). Computers in Human Behavior, 26(1), 100–109 doi.org
- VRChat Safety and Trust System docs.vrchat.com
- Digital body, identity and privacy in social virtual reality: A systematic review. Frontiers in Virtual Reality (2022) doi.org
- Gender expression and gender identity in virtual reality: avatars, role-adoption, and social interaction in VRChat. Frontiers in Virtual Reality (2024) doi.org
- Son, G., & Rubo, M. (2025). International Journal of Human-Computer Studies, 199, 103488 doi.org
- Lapidot-Lefler, N., & Barak, A. (2012). Computers in Human Behavior, 28(2), 434–443 doi.org
- Duronto, P. M., Nishida, T., & Nakayama, S. (2005). International Journal of Intercultural Relations, 29(5), 549–560 doi.org
