Cognitive behavioral therapy (CBT) is one of the most common types of talk therapy in the United States. There are 11 criteria by which cognitive-behavioral therapists-in-training are typically judged. What if your skills could be assessed and improved with AI feedback?
It is the first study of CBT sessions conducted with real people in real therapeutic conversations. The findings have recently been published in PLOS One .
Evolving psychologists
More than 1,100 actual conversations between therapists-in-training and patients were analyzed by an AI created by the Signal Analysis and Interpretation Laboratory (SAIL) at the Viterbi School of Engineering at the University of Southern California. For therapists who are learning, it is other human evaluators who will normally evaluate their sessions. But this AI was able to match what a human evaluator could achieve with 73 percent accuracy .
The AI could thus judge the interpersonal skills of the therapist and discern if the therapists created the appropriate structure for the session (if they tackled the task assigned to the patient, for example). In addition, the AI could tell if a therapist was adequately patient-focused rather than sharing too much of his own story and if he was able to collaborate with his patient and establish a relationship. All these aspects are taken into account to generate a single aggregate quality metric .
Such evaluations, typically performed by humans, are necessary to train and provide performance-based feedback to a therapist, leading to better clinical outcomes. The goal, the researchers say, is to automatically generate metrics from a recorded session to facilitate these applications. The next step is to add tonal or prosodic qualities of spoken interactions to this tool to enrich its capabilities.