Empathy is the ability to connect with others’ feelings and experiences, which can be incredibly beneficial in building strong relationships and excelling in certain job environments. This has led researchers in psychology to focus on creating reliable ways to assess empathy.

Traditionally, empathy measurements have relied heavily on self-assessments and surveys like the interpersonal reactivity index (IRI) or the Empathy Quotient (EQ) test. However, recent efforts have explored innovative methods employing machine learning and other computational techniques.

A team from Hong Kong Polytechnic University has pioneered a novel machine learning-based video analysis tool that aims to evaluate the empathy of individuals captured on film. Their framework, detailed in a preprint published in SSRN, could be valuable for research in organizational psychology and various empathy-related fields.

“Our work is inspired by the ongoing AI revolution, which Huang and Rust (2021) describe as the onset of a new ‘feeling economy,’ where machines take on analytical tasks while humans focus on emotional and relational dynamics,” Li Cui, one of the paper’s authors, shared with Tech Xplore.

“In this landscape, workplace empathy is more crucial than ever. Yet, many leadership studies mainly highlight negative traits such as overconfidence and narcissism, giving a one-sided view. We see an opportunity here: AI-driven video analytics can provide direct and scalable insights into human traits, moving beyond outdated and biased measurement techniques.”

“We created a machine learning framework that identifies emotional mimicry in CEO interviews, using this as an indicator of empathy,” Cui explained. “This framework allows us to measure empathy in a convergent manner, validated against multiple signals in real-world contexts. By capturing intricate behavioral details, it offers an efficient and scalable method for empathy assessment.”

The researchers employed their innovative approach to analyze how CEOs exhibit empathy during television interviews, where they answered questions about their companies. Their findings suggest that this framework is a promising tool for gauging CEO empathy.

“Traditionally, measuring empathy at scale has been a challenge. Our framework captures subtle behavioral signals, such as emotional mimicry, and validates them in real-world scenarios,” Cui stated. “Our data indicates that more empathetic CEOs lead to better corporate policies, such as decreased workplace injuries and fairer compensation practices.”

“Moreover, empathetic leaders seem to manage crises more effectively, which could enhance company value. These initial findings highlight empathy’s potential significance as a leadership quality, representing a step forward in how we study soft skills in real-life situations.”

The researchers aspire to enhance their framework, further contributing to tools in FinTech and psychology for studying traits, communication habits, and behaviors in both work and daily settings. In the future, it could also enhance AI analytics platforms, allowing for predictions of empathy based on body language.

“Our current research is just the beginning of exploring how video analytics can quantify soft traits like empathy, with plenty of potential for further growth,” Cui remarked. “One future endeavor could be applying this framework to a wider array of video contexts—like social media clips or cross-cultural settings—to broaden its applicability and understand how empathy manifests differently.”

As they continue their research, the team plans to refine their framework to minimize biases in interpreting nonverbal cues. Additionally, they aim to create similar frameworks for assessing other human traits such as assertiveness, emotional stability, and trustworthiness.