This is my reflection on Enhancing Our Lives with Immersive Virtual Reality by Mel Slater and Maria V. Sanchez-Vives (Frontiers in Robotics and AI, 2016). It’s an opinion paper that reviews VR applications with some scientific support. Its main argument is that VR works both as a “reality simulator” and as an “unreality simulator”, because people respond realistically even to situations that could never happen.
What I liked #
The strongest use case for me is research that is no longer ethically possible. Milgram’s obedience studies relied on deception, which led to today’s rules on informed consent and the right to withdraw. The VR version of the experiment by Slater et al. (2006) got ethics approval because participants knew the Learner was virtual and could stop at any time. The obedience pattern still appeared, with stress measured through skin conductance (how much the skin sweats) and ECG (heart activity). The split between conditions is telling: all 11 participants who only communicated with the Learner through text gave every shock, while 6 of the 23 who could see and hear her withdrew. To me this is plausibility illusion (Psi) in practice.
The second point is the gap between moral judgment and moral behavior. In one desktop VR study, participants rated sacrificing one person to save five as less acceptable when the dilemma was described in text, but were more likely to do it in VR. The paper reports the same tendency in the other trolley-type studies. A questionnaire captures what people think they would do. VR gets much closer to what they actually do.
I also liked the presence paradox. The paper separates immersion, the technical capability of the system, from presence, the subjective response to it (place illusion plus plausibility illusion). Presence includes surprise: you know you are at home wearing a VR headset, but you feel as if you are in the Himalayas. In physical reality there is no such surprise. So if immersion became perfect, presence would disappear. In the authors’ words, presence “represents the nonperfection of immersion”. It turns the usual assumption around, since better technology does not always mean more presence. The authors add that a generation growing up with VR may also lose this surprise, simply because they get used to it.
What I did not like / what caught my attention #
The scope is one-sided. I understand that the title is about enhancing our lives, so focusing on positive uses is a logical choice. But in a 47-page paper, harmful uses get a single sentence saying some will be “morally repugnant”. Privacy and addiction are not discussed. Simulator sickness only appears as a side note inside individual studies. The same mechanisms the paper praises, such as body ownership changing implicit attitudes, could also be used to manipulate people. Given the length, I would expect at least one paragraph on the risks.
The racial bias result is the claim I trust least. Peck et al. (2013) report that 12 minutes of embodiment in a Black virtual body reduced implicit racial bias, with the effect still present after a week (Banakou et al., 2016). The measure is the Implicit Association Test (IAT), which the paper admits “may be controversial”. It also notes that an earlier study (Groom et al., 2009) found the opposite. Both studies are co-authored by Slater. One independent rubber hand study (Maister et al., 2013) points the same way, but I would want an independent replication of the full body result.
Finally, the ethical advantage of VR is also its validity problem. When participants in the VR Milgram study were asked why they continued despite feeling uncomfortable, a typical answer was “I kept reminding myself that it wasn’t real”. Their stress was also lower than in the original. The authors admit that knowing it is not real “can also dampen down” responses. So VR sits somewhere between a questionnaire and reality. How big that gap is probably depends on the scenario.
My take #
Across the whole paper, from surgical training and sports to cultural heritage and immersive journalism, every application relies on the same idea: people respond to VR as if it were real, even though they know it is not. The most convincing results are the ones checked against real-world outcomes. In Seymour et al. (2002), surgeons trained in VR were 29% faster in real gallbladder surgery and six times less likely to make errors. That system was actually a 2D screen, not an immersive headset, which connects to the authors’ own complaint in the conclusion that many papers call a normal PC setup “VR”. The weakest results rely on indirect measures like the IAT and come mainly from the authors’ own group. So I see VR as a strong tool for research and training, but its claims need the same checking as any other method.
Reference: Slater, M. and Sanchez-Vives, M. V. (2016). Enhancing Our Lives with Immersive Virtual Reality. Frontiers in Robotics and AI, 3:74. https://doi.org/10.3389/frobt.2016.00074
