In a new paper published in Science Robotics, a cognitive roboticist, cognitive psychologist and a psychiatrist discuss the concept of “sense of self” in humans, and they explore how robots can be used to better understand the phenomenon.
The intrinsic experience of having a self—that feeling of being contained within our bodies and actively engaging with the world—is a fundamental aspect of human existence. We naturally perceive ourselves as enduring selves from day to day and recognize others as having their own selves as well.
The authors propose that robots can play a dual role: acting as embodied models of self and serving as innovative testing platforms for psychological research. By engineering robots to exhibit processes that contribute to the human “sense of self,” we can unlock new insights into this complex phenomenon.
The researchers are Agnieszka Wykowska, head of the Social Cognition in Human-Robot Interaction unit at the Istituto Italiano di Tecnologia (IIT) in Italy; Tony Prescott, Professor of Cognitive Robotics at the University of Sheffield in the UK; and Kai Vogeley, Professor at the Department of Psychiatry and Psychotherapy at the University of Cologne in Germany.
The inquiry highlights a profound reality: the human sense of self is deeply intertwined with our physical bodies and the dynamics of our actions and interactions. Cutting-edge research in human cognition reveals that the human sense of self comprises multiple ongoing processes, including the sensation of “owning” a body and the feeling of having “agency,” which refers to control over one’s actions.
Currently, roboticists aim to create robots that can effectively recognize their own bodies (the self-other distinction) and understand the outcomes of their own actions (agency). From this viewpoint, robots can act as embodied models representing the cognitive processes in humans that inform the sense of self.
Moreover, robots can also serve as experimental tools for investigating the sense of self, given that they have bodies and can engage with both humans and their surroundings. The three authors discuss utilizing robots in these two distinct ways. The first approach involves programming robots to mimic processes within the human mind and brain linked to the experience of self, as interpreted through psychology and neuroscience.
Existing research indicates that the sense of self in humans develops as the brain posits the best explanation for its sensory experiences and its function in generating those sensory signals. As a physically embodied entity, a robot presents a fitting platform for testing these theories.
The second approach emphasizes the use of robots in groundbreaking psychological experiments that explore human interaction with machines exhibiting social capabilities, such as language communication and joint attention. These studies are essential for understanding whether individuals truly perceive these robots as social entities and if their mental frameworks when engaging with robots mirror those experienced during interactions with other humans.
The research from Wykowska’s team at IIT has demonstrated that humans can sometimes forge a sense of joint agency with robots, especially when collaborating as a team and treating robots as intentional agents. In addition, the authors draw intriguing connections between the evolution of self-concept in humans and the potential for replicating similar features in robots.
By around age four, children develop an understanding of their continuous existence and recognize that others possess sensations of self, too. This important aspect is now being methodically investigated in robotics through the development of memory systems designed to mimic human autobiographical memory. However, it is crucial to note that this field is still in its early stages; current robots lack the self-awareness to persist from day to day and do not perceive other entities, whether human or robotic, as having independent selves.
Researchers are starting to explore these self-related aspects in robots by developing memory systems that resemble human autobiographical memory. Nevertheless, this line of research is still in its infancy; at present, robots lack an understanding of their own persistence over time and do not recognize others (human or robotic) as having selves.
Additionally, the article underscores potential future research avenues and existing challenges in comprehending self-awareness through robotics, particularly when it is affected in individuals due to certain conditions like schizophrenia or autism. By exploring this diversity, the authors hope to offer scientists new perspectives on the fundamental components that contribute to the experience of self.
Journal reference:
- Tony J. Prescott, Kai Vogeley, Agnieszka Wykowska. Understanding the sense of self through robotics. Science Robotics, 2024; DOI: 10.1126/scirobotics.adn2733