Skip to Content

Virtual humans, explained

AI characters that talk with people, and how to use one without losing trust.


Talk to Virtual Terence

What a virtual human is

USC's Institute for Creative Technologies, one of the oldest labs in the field, describes virtual humans as autonomous agents that support face-to-face interaction with people in virtual environments.

In practice that means a character, sometimes modeled on a real person, that can listen, answer and react. It might live on a website, in a training simulation, in a video or inside a headset.

Why people open up to them

ICT researchers found that people who believed they were talking to a fully automated virtual interviewer disclosed more than people who believed a human was operating it. That is powerful, and it is exactly why honesty and disclosure matter.

Training

Practice hard conversations, from interviews to customer service, with a character that gives structured feedback.

Service

Answer the questions every visitor asks, in your voice, at 2 a.m., and hand the real conversations to you.

Storytelling

Bring a brand, a founder or a historical figure into a scene people can talk with, not just watch.

Doing it right

  • Always disclose. Say plainly that it is AI. California's BOT Act already requires bots that sell or try to influence a vote to say so.
  • Use only your own likeness, with consent. California law now restricts digital replicas of performers and of people who have died without permission.
  • Never fake an endorsement. The FTC's rule bans AI-generated fake reviews and testimonials.
  • Check how it renders you. Research going back to MIT's Gender Shades study shows face technology has performed worse on darker skin. Test your avatar under different lighting before you publish it.

Want the same idea applied to a place instead of a person? That's a digital twin, and you can walk through one.

Meet Virtual Terence

This is a working example of a virtual human: an AI version of me that answers the way I would in a first call, and says plainly that it is AI. Ask it something real. When it reaches the edge of what it knows, it will point you to the right page or to the real me.

Terence Latimer
Virtual Terence AIAn AI version of Terence, trained on how he answers first-call questions. It can be wrong, and it is not the real person.
Chats are saved so Terence can improve the answers. Don't share private details. Want the real one? Book a free consultation.

Want a virtual version of your expertise?

We'll map what it should know, what it must never say, and where it hands off to you.