Human Interaction Enhances Digital Mental Health Tools, Research Indicates

A significant portion of the global population, approximately one in eight individuals, contends with mental health disorders; however, fewer than 50% receive adequate treatment. A recent study conducted by researchers at the University of Reading, published in PLOS One, demonstrates that incorporating a human element--even in scripted interactions--substantially improves emotional engagement and the perception of empathy during online psychological assessments.

The study involved 75 participants who underwent three types of brief online interviews focused on their well-being. These formats included one with a semi-scripted live interviewer, another featuring a live but fully scripted interviewer, and a third employing fully automated pre-recorded videos. Results indicated that participants perceived both live interviewers as significantly more empathetic compared to the automated option, while facial recognition technology revealed heightened expressions of joy during live interactions.

As mental health services continue to adapt, this research suggests that hybrid models, which integrate automation with human contact, may represent the most effective approach moving forward. Furthermore, the findings provide valuable insights for future developments in artificial intelligence.

Dr. Thomas Nyman, the lead researcher, noted that even when interviewers adhered strictly to a script, the mere presence of a human interviewer made a substantial difference in participants' experiences. With advancements in AI technology, there is potential to create digital agents capable of mimicking the human qualities that participants responded positively to in this research, such as empathy, adaptability, and presence. Recognizing the emotional needs of individuals is crucial for developing technology that meets those requirements in a human-like manner.

For further details, refer to the article titled Emotional engagement and perceived empathy in live vs. automated psychological interviews published in PLOS One (2025). DOI: 10.1371/journal.pone.0323490