An interdisciplinary team from the University of Southern California has developed an automated system for evaluating psychotherapists though signal processing. Their work blends psychology with electrical engineering and computer science, with the common goal of measuring the effectiveness of psychotherapy. The protype system they created assesses empathy from audio recordings and quantifies the level of empathy. The system automatically detects an “empathetic opportunity”—a situation when a patient expresses an emotion. Then computers evaluate whether the therapist responds, and if so, whether the response is empathetic. Certain phrases are considered more empathetic, but it’s not just about what the therpaists say, it’s also how they say it. Loud and high-pitch voices, for example, are associated with lower empathy. Even laughter is measured and analyzed as a non-verbal cue that signals an important part of a conversation.
This new method of evaluating therapists has the benefit of objectivity because it eliminates the human bias. However, it’s hard to say for certain that algorithms can measure something as abstract as empathy. “This is very human-centered engineering,” one researcher said. “The role of technology is to support, not to supplant human expertise and intelligence.”