Skip to main content

Study flaws might be obscuring the role of personality in human-robot interactions


During the pandemic, robots are becoming an important way to deliver health care. But there remains a lack of understanding about the role personality plays in interactions between these robots and patients. That’s according to a new meta-analysis from researchers at the University of Michigan, who found “several gaps worthy of attention” in over a dozen studies on the subject of human-robot interaction in health care contexts.

Even before the health crises, the demand for health care services was expected to outpace the availability of health workers. Robots might help to fill the gap, but a lack of understanding around personality threatens to create barriers to adoption for health care systems. Without consensus, it’s difficult to know which personality traits a robot should or shouldn’t have, for example — or indeed if personality matters to patients at all.

The researchers canvassed 1,069 studies and focused on 18 published between 2014 and 2018 in which a total of 805 people participated. (They eliminated the rest according to an eligibility criterion through several rounds of screening.) Around 39% of the papers investigated how participants’ personalities affected their interactions with a robot, while 44% explored how robots’ personalities affected a person’s interactions. The remaining 17% looked at how a participant’s and a robot’s personalities worked together to create the person’s experiences with the robot.

Problematically, the average sample size across the 18 studies was small (44.7), which researchers note can lead to a propensity for errors. (The mean sample size excluding three large-scale studies averaged only 23.8.) Moreover, there was “significant” variation with respect to participants’ genders — two-thirds had gender imbalances — indicating the results are skewed. And people from one region — Europe — were most frequently sampled with 64% of reported samples, to the exclusion of those from South America and Central America.

VB Event

The AI Impact Tour – Atlanta

Continuing our tour, we’re headed to Atlanta for the AI Impact Tour stop on April 10th. This exclusive, invite-only event, in partnership with Microsoft, will feature discussions on how generative AI is transforming the security workforce. Space is limited, so request an invite today.
Request an invite

These weren’t the only problems the meta-analysis identified. None of the studies recruited participants over 65 in assisted living settings and not one investigated personality in group interactions (e.g., among a team of health care workers). Only two tried to tease out the interplay between humans’ and robots’ personalities, which could provide valuable design insights. And the studies most often drew on the “big five” traits — extraversion, agreeableness, openness, conscientiousness, and neuroticism — as measures of personality, rarely examining traits like helpfulness, reliability, intelligence, and confidence that have been shown in previous research to help physicians deliver quality care.

It’s the coauthors’ belief that future research is needed to expand the health care robotics field’s understanding of personality. “The use of robots as health care providers is only expected to increase. This necessitates a need to reflect on what has been done in this area and to contemplate what still needs to be done,” they wrote. “Robots are becoming an important way to deliver health care across the world, and personality is vital to understanding their effectiveness. This paper is an important starting point in establishing an understanding of personality in human-robot health care interactions.”

VB Daily - get the latest in your inbox

Thanks for subscribing. Check out more VB newsletters here.

An error occured.