This NPR story alerted to the newest kind of psychological therapist being built by the the University of Southern California Institute for Creative Technologies (ICT), in conjunction with the Pentagon. I say built because “Ellie,” the experimental therapist, is not a person, but rather a computer. In this instance “she” is being used to interview veterans and screen them for signs of depression and post-traumatic stress disorder.
Ellie works by asking the research subject questions and recording/analyzing his or her responses, as shown in this demonstration video:
Ellie focuses not on what the interviewee says, but on his or her non-verbal cues, using an interactive technology called multisense:
Multisense automatically tracks and analyzes in real-time facial expressions, body posture, acoustic features, linguistic patterns and higher-level behavior descriptors (e.g. attention, fidgeting). From these signals and behaviors, indicators of psychological distress are inferred to inform directly the healthcare provider or the virtual human.
Studies have shown that Ellie interprets some of these expressions quite well. Reportedly she asses the genuineness of a smile, for example, as well as a human psychologist.
This is a nifty technological achievement, certainly. But what’s just as interesting to me as the computer capability is people’s willingness to interact with the computer. Many people not only carry on the somewhat awkward conversation with the machine, but they actually share more than they would with a human counterpart:
“People opened up more to the virtual human than to a real person. They said they felt less judged by the virtual human,” [ICT psychologist Gale] Lucas said, even though they knew that their answers would be recorded and viewed later by workers in a lab.
“It’s about what’s happening in the moment — having a safe place to talk,” she added.
Why are interviewees more open when talking to a computer? An article in Computers in Human Behavior elaborates:
participants who believed they were interacting with a computer reported lower fear of self-disclosure, lower impression management, displayed their sadness more intensely, and were rated by observers as more willing to disclose. These results suggest that automated VHs [virtual humans] can help overcome a significant barrier to obtaining truthful patient information.
Lower “impression management”? That is the “goal-directed conscious or subconscious process in which people attempt to influence the perceptions of other people . . . they do so by regulating and controlling information in social interaction.” In other words, interviewees are less concerned about being judged by the machine, and express themselves in a less inhibited manner.
Computers can also facilitate communication by being better listeners. They can be programmed not to interrupt the speaker. And they avoid distracting body language–e.g., looking around, checking the phone, and yawning. Today it’s challenging finding a human who gives such undivided attention.
At any rate, given continuous technological advancements, we’re sure to chat with more virtual humans in the future.