Social robots could help diagnose and treat autistic children
For the past four years, Brian Scassellati has been researching the uses of humanoid robots that can interact with people by natural social cues. Scassellati spoke last Friday on ?Social Robots, Social Development and Social Disorders,? in which he demonstrated the uses of social robots when applied to autism diagnosis and therapy.
Scassellati, a computer science professor at Yale, developed a social robot named Kismet. Kismet has a physical body and is autonomous.
?The temporal correlations between motor activity and perceptual input discriminate [the robot] self from other time constants determined autonomously,? Scassellati said. In this way, Kismet can distinguish itself from other things. ?A lot of philosophers argue that a sense of self is the basis of consciousness,? said Piper Lincoln, a sophomore ethics and history major. ?It?s pretty startling to think of the implications of this research.?
Kismet has auditory, visual, and expressive systems that enable it to participate in human-social interaction and to demonstrate simulated human emotion. It reciprocates emotion through facial expressions such as moving its ears, eyebrows, eyelids, lips, jaw, and head. Scassellati?s goal is to use these robots to teach social skills to children with autism.
Autism is a brain disorder that interferes with a person?s ability to communicate and relate to others. A child with autism fails to respond or interact like other children of the same age.
Robots can help diagnose autism with ?hard-core results,? Scassellati said, by documenting how the child reacts in a social context, using speech and facial expressions. For example, a typical child loses interest in a generic robot; he or she prefers interacting with a face. However, an autistic child has no preference for a regular face over a scrambled face and does not lose interest at all.
Scassellati?s robots are programmed to recognize social cues specific to autism and can provide Scassellati with quantitative information that can be used by physicians to diagnose, track the progress of a patient, and make comparisons between patients within a population.
Though Kismet seems like a step forward in medical technology, others are skeptical. ?I?m sort of afraid of how much the kids will be treated like science projects,? said Phoebe Wu, a senior information systems major. ?Because when you?re using these robots to collect data ... what else can you do? You don?t want the focus to shift from helping kids to getting more data.?
Scassellati said that though his results as of now seem solid, he used a relatively low number of subjects. Therefore, he does not want to give false hope to families with autistic children. His research is preliminary, and there are still many more variables to consider.