The recent surge of virtual reality (VR) headsets has generated daily conversations around the possibilities and the implications of VR experiences. One of the most intriguing aspects of immersive VR experiences is the fact that the virtual body does not have to mirror the physical body but can be represented with limitless digital creations.
Jeremy Bailenson, the founding director of the Virtual Human Interaction Lab (VHIL) and a professor in the department of Communication at Stanford University has been a pioneer in the area of virtual reality. His main area of interest is digital human representation, especially in the context of immersive virtual reality. We reached out to J. Bailenson and A. Stevenson Won, a PhD candidate, to talk about VHIL and recent research at the lab.
Lilia Ziamou: What are the goals of the VHIL? What is the main focus of your research program?
Jeremy Bailenson: Our mission at VHIL is to understand the dynamics and implications of interactions among people in immersive virtual reality simulations, as well as other forms of human digital representations in media and communications systems. We focus on using empirical, behavioral science methodologies to explore people as they interact in digital worlds. Our research programs tend to fall into one of three larger categories:
- What new social issues arise from the use of immersive VR communication systems?
Lilia Ziamou: In earlier conversations, you have highlighted the fact that in VR environments there are no rules. What are the consequences of this?
Jeremy Bailenson: Among other advantages, virtual environments give us the opportunity to embody avatars in completely novel ways. Avatars are real time digital representations of the user in virtual reality; however, unlike real life, these avatars don't need to resemble the real person. In VR we can customize avatars to perform better at tasks, to communicate more effectively, or to impact social norms.
Lilia Ziamou: The traditional approach to avatar design has favored biological representation. What do we know about avatars that are not constrained by biological representation?
Andrea Stevenson Won: When we think of a VR experience, we often think of it as following the rules of the real world. For example, people can meet in VR using avatars and talk "face to face" just like they do in the physical world. But it doesn't have to be that way. For example, if I take one step in real life, why should I take just one step in the virtual world? Instead, I can fly or swim through the virtual environment by moving my arms. Similarly, when you think about manipulating virtual objects, there is not necessarily a reason why we should be constrained to using two hands. So, for many applications, we may want the avatars in a virtual environment to look and move like real people- but not always.
Lilia Ziamou: Recent research at VHIL shows that the design of novel avatars can influence users' engagement. Tell us about this research.
Andrea Stevenson Won: VR allows people to inhabit -and control-novel bodies, that is to say, avatar bodies that aren't like normal human bodies at all; for example, avatars with three arms. Jaron Lanier called this concept "homuncular flexibility". In our previous work, J.N. Bailenson, J.D. Lee, J. Lanier and I showed that people can quickly learn to control such novel avatars when those novel avatars are better suited to a given task.
The next question was, what are the possible effects of avatar appearance on task success? In other words, when you have an avatar that doesn't move like a real person, how should it look? We tested how the avatar's appearance alone might affect people's success in a target-hitting task. Everyone in the experiment controlled an avatar with three "arms" by moving their left and right arms and rotating their wrists in the real world, but we varied how the avatar appeared. Some people controlled avatars with that third extension shaped like a human arm and matched to the user's skin tone. Other people controlled a more 'mechanical' appearing avatar, where the third 'limb' was a simple silver geometric shape. We also varied whether the extension was "floating" in front of the users, or whether it appeared to be attached to their avatar's chest.
Interestingly, people who controlled 'biological' appearing avatars were somewhat less successful at a target-hitting task, when the extension was floating in front of their avatars' bodies. So, although we often think that more realistic avatars are better, this may not be the case when the avatar moves in a way that isn't what we are used to.
Lilia Ziamou: What are you currently looking at?
Andrea Stevenson Won: There's a lot more to explore regarding avatar design! Bireswar Laha, a post-doc at VHIL, and I were just discussing how to make novel avatar bodies more intuitive to control. One approach is to look at how people move naturally and the natural constraints of the body, and try to push movements in a different direction. For example, instead of controlling a "third arm" by using your left and right wrists, it might feel more natural to control it by just rotating one wrist.
We also have cultural associations of what movements mean. Different movements may have positive or negative associations. For example, a recent study at VHIL looked at the effects of embodying a superhero while in VR. In this experiment, the authors found that people who "flew" around a virtual city by holding their hands over their head like Superman were more helpful later than people who "flew" around the city in a helicopter. So, we should consider what associations come with gestures.
As virtual reality becomes more widely used, it's increasingly important to understand how avatars affect their users. We can also consider how avatars can be designed to potentially change the way people behave in virtual and in real worlds.