Brave New World: AI and interviews
In the workplace, psychometric tests are used along with traditional interviews as techniques in identifying potential hires. Established psychometric tests sum the scores of individual indexes or scales (ex. WAIS-IV: VCI, PRI, WMI, PSI) which each represent a different dimension of personality or ability. The simplicity of this approach is what has allowed for the interview to still be held as essential. There are certain complexities in the interview which a test, ultimately, cannot catch. But is this for certain?
With the advancement of artificial intelligence techniques, such as expert systems and neural networks, has come the opportunity for improved psychometric testing. As technology advances, can AI replace the in-person interview? Now it is possible using neural network technology to train computer systems to detect patterns in responses to questionnaires. Neural networks offer a predictive power which cannot be matched by classical psychometric tests.
It has been shown that one of the limits of the existing psychometric test is its tendency to produce “clone workers”. If a particular personality profile is identified as ideal, say the ideal salesman, you end up with a team or company of ideal yet very similar workers. This ignores the importance of diversity in the workplace. We know in reality there are many different types of effective salesman, each with different means to the similar ends. Neural networks able to recognize the myriad of different, non-linear pathways to success have the potential to outperform any traditional linear technique which focuses on the simple trait. AI techniques like neural networks could represent an information boon for HR professionals in the workplace. The gap is closing. But what about AI technology in the therapeutic space?
Meet Ellie. She’s in session with a veteran. She listens to his voice — for his tone and his pitch. She looks for meaning in his posture. She grimaces as she hears of his trauma. Ellie, it seems, is good at her job.
Ellie is an avatar, a virtual therapist engineered at USC’s Institute for Creative Technologies with funding from DARPA, the US Defense of Department’s advanced research center. Ellie is capable of reading and responding to human emotion in real time. But why would anyone choose a virtual therapist over a real one?
When faced with revealing difficult, often traumatic experiences, people often don’t give doctors the full story. Yet the researchers behind Ellie suspected that people may be more willing to talk if presented with an avatar. So they put it to a test– 239 people were either told Ellie was an AI or that Ellie was being controlled remotely by a therapist. Those that thought Ellie was under somebody’s control were much more reticent in disclosing personal information. Those dealing with the AI Ellie felt less judged, and more willing to be open and honest. It makes sense why the DOD would be interested in virtual therapy for its soldiers. Soldiers so very often find it difficult to talk about their trauma. For soldiers with PTSD, something like Ellie could facilitate the dialogue necessary for treatment.
The potential applications for AI across psychology seem boundless. The future is now — the AI revolution is on.
