In this interview, Professor Stefano Guidi from the DISPOC of University of Siena reflects on the role of immersive technologies such as virtual and extended reality in industrial training. Through the EMPAIRED project, opportunities and challenges emerge related to procedural learning, interaction with artificial agents, and the social perception of intelligent technologies.
The video interview was carried out by Santa Chiara FabLab from the University of Siena. Below the video, the full transcript is given.
Interviewer: Good morning everyone. Today we are here with Stefano Guidi, Professor of Social Psychology at the Department of Social, Political and Cognitive Sciences at the University of Siena. Thank you for accepting our invitation.
Prof. Guidi: Good morning, thank you.
Interviewer: To begin with, could you explain what you work on and what your research interests are?
Prof. Guidi: Yes, certainly. I am an Associate Professor at the Department of Social, Political and Cognitive Sciences at the University of Siena, where I teach several courses, mainly in psychology. I teach courses on the foundations of psychology, social psychology, and also research methods in psychology.
As for my research interests, they are quite broad, partly as a reflection of my academic path, which has not been entirely linear, let’s say. In fact, I followed a trajectory that started from cognitive psychology, where I worked both on visual perception and on the application of psychological knowledge about cognitive processes to design—particularly in the field of cognitive ergonomics. Over the years, I gradually shifted more toward conducting research and evaluation practices in the field of work psychology and later social psychology.
Currently, the topics I work on concern, on the one hand, stereotypes and prejudices related to human beings—therefore topics that are clearly within social psychology—and, on the other hand, the perception of new technologies such as robotic systems and artificial intelligence systems.
Interviewer: Today we are here to speak with you about EMPAIRED, a project that we coordinate and are developing together with two other organizations: Moebeus and Deep Reality.
EMPAIRED is one of the projects selected as a winner of the Open Call 2 of MASTER XR, a European project that promotes the adoption of Extended Reality technologies in industrial training and manufacturing robotics through an open and collaborative platform. The EMPAIRED project, specifically, aims to develop an empathetic ecosystem based on the integration of virtual reality and supported by a virtual assistant from the Algho platform. The ErgonXR platform (developed during Open Call 1 of the MASTER project) is instead responsible for the real-time monitoring of physical ergonomics data.
We will now show you a video of the scenario we are currently developing.
DEMO
After watching the demo of the EMPAIRED project, what do you think could be the potential benefits and possible limitations of using VR and XR technologies for training in industrial contexts?
Prof. Guidi: Virtual reality certainly provides scenarios in which interactions with the environment can be simulated with a high level of realism and in a safe way for the person who is performing those interactions, that is, for the user interacting with the environment.
This clearly opens up important opportunities for training related to technical procedures and offers the possibility of experiencing firsthand the effects of risky behaviors. This is because when we perform certain tasks, the way we learn is fundamentally based on specific memory systems: procedural memory, which strongly involves the body. In this case, learning takes place through the internalization of control over how the body, and our bodily movements, must react and respond to environmental stimuli.
When we learn a procedure, we learn to respond more and more automatically to stimuli coming from the environment, which we progressively internalize until we reach an automated form of control. Clearly, a tool such as virtual reality—which makes it possible to simulate environmental stimuli with great fidelity, to track movements with precision, to monitor them in real time, and to perform them safely—provides an optimal environment for this type of training related to procedures.
In fact, it is no coincidence that simulators—that is, forms of virtual reality, even if somewhat different from this one—have long been used for training. For example, airplane pilots and Formula 1 drivers use simulators precisely because of their realism, as they allow them to experience this type of control, to practice risky procedures, and to refine their control over movements. In surgery as well, virtual reality is often used to simulate and learn these procedures without risks—in these cases more for the patient than for the user.
The main limitations that I see, perhaps—at least in this type of environment, as shown in the video you presented—are mainly related to input devices, such as the controllers (I noticed the joypads), which are the tools through which the user interacts with the virtual environment. In my opinion, these are still not sufficiently realistic and may produce certain distortions or artificial movements that would not occur in reality and that could therefore, in some way, interfere with the learning of the procedure.
Interviewer: In your opinion, what characteristics should simulations in virtual environments have in order to support the learning of ergonomic safety rules in real working contexts?
Prof. Guidi: Certainly, realism and a strong sense of immersion are, in my opinion, the most important factors in facilitating the transfer of what is learned in a virtual environment to a real one. They are also important in ensuring that users take the simulations seriously and engage with them with the necessary level of commitment—which is important from a motivational perspective in order to continue the activity—and therefore proceed with the exercises, which are ultimately what lead to the learning of these types of procedures.
Interviewer: In one of your recent publications — the paper “Unfair game: how age and robot deception shape the attribution of mental states in virtual reality” — you discuss how the perception of artificial intelligence changes depending on the application context. In your opinion, what advantages could arise from integrating embodied chatbots, used as coaches, within virtual environments?
Prof. Guidi: First of all, I should point out that, in the study you are referring to, my colleagues and I mainly focused on non-embodied AI systems—that is, systems presented as applications designed for different types of tasks across various application domains. Our results showed that people nevertheless tended to trust and accept the use of these systems more in rather technical domains, such as insurance and healthcare, compared to more relational domains. This seems to suggest that, even for technical training such as that carried out in industrial contexts, these types of AI systems could be well accepted and perceived positively.
Anthropomorphism, and the use of embodied chatbots such as avatars, should increase—and could further facilitate—this kind of acceptance. However, it is also important to consider some possible obstacles or barriers. For example, there is the phenomenon known as the Uncanny Valley, which occurs when an artificial agent appears very similar to a human being, while at the same time people clearly recognize that it is not human. This can generate a negative feeling or a negative attitude toward that type of agent. Therefore, further studies would probably be needed to better understand when, and under what conditions, an avatar might trigger this type of negative response.
Another aspect that comes to mind, and to which attention should be paid, concerns the effects related to the gender of the assistant. We know from other studies—for example in social robotics—that the gender of robots can actually activate gender stereotypes that originate in human social perception. These stereotypes could therefore hinder both the interaction with and the acceptability of these chatbots, which might otherwise be quite well accepted in this technical domain, in my opinion.
Interviewer: With this question we conclude our interview. We thank you once again for the time you dedicated to us and for your valuable contribution.
Prof. Guidi: Thank you.
To stay informed, leave your contact details here.