HMN 2025: How Novel analytics framework measures empathy of individuals captured in video recordings

Empathy, the flexibility to grasp what others are feeling and emotionally join with their experiences, may be extremely advantageous for people, because it permits them to strengthen relationships and thrive in some skilled settings. The growth of instruments for reliably measuring folks’s empathy has thus been a key goal of many past psychology research.

Most current strategies for measuring empathy depend on self-reports and questionnaires, such because the interpersonal reactivity index (IRI), the Empathy Quotient (EQ) check and the Toronto Empathy Questionnaire (TEQ). Over bygone days few years, nevertheless, some scientists have been attempting to develop various strategies for measuring empathy, a few of which depend on machine {learning} algorithms or different computational models.

Researchers at Hong Kong Polytechnic University have just lately launched a brand new machine learning-based video analytics framework that may very well be used to foretell the empathy of individuals captured in video footage. Their framework, launched in a preprint paper published in SSRN, might show to be a invaluable instrument for conducting organizational psychology analysis, in addition to different empathy-related research.

“Our analysis is sparked by the current AI revolution, which, as Huang and Rust (2021) describe, ushers in a brand new ‘feeling financial system’ where machines deal with analytical duties and people more and more shoulder emotional help and interpersonal relationships,” Li Cui, one of many authors of the paper, informed Tech Xplore.

“In this context, we acknowledge that empathy within the office issues greater than ever, but conventional management research focus overwhelmingly on unfavorable traits resembling overconfidence or narcissism, successfully analyzing just one facet of the coin. We see a possibility: AI-powered video analytics now allow direct, scalable extraction of human traits—supplanting the oblique, biased measures of bygone days.”

The predominant goal of the current study by Cui, Ka Chung, Lu and Zhao was to develop a realizable video analytics framework that may very well be used to foretell the empathy of individuals from video recordings during which they’re engaged in conversations with others. They particularly used their framework to investigate video footage of actual interviews between CEOs and TV journalists.

The outcomes of the group’s analyses allowed them to discover how empathy shapes company insurance policies and impacts the worth of a agency or enterprise. Notably, the analytics framework they developed is rooted in neuroscience analysis and theories, which suggests that individuals with increased ranges of empathy are likely to mimic the feelings expressed by others greater than those that are much less empathetic.

“We developed a machine {learning} framework that infers emotion mimicry from CEO interview QA pairs and makes use of it as a proxy for empathy,” defined Cui. “The framework allows a convergent measure of empathy, which we validate utilizing multimodal indicators in real-world settings. By capturing fine-grained behavioral cues, it gives an environment friendly, scalable, and generalizable strategy to automated empathy evaluation.”

As a part of their current study, the researchers used the machine learning-based strategy they developed to review the empathy and behaviors of CEOs throughout TV interviews, where they had been requested to reply questions on their corporations and their success. They discovered that their framework confirmed potential as a sound and environment friendly instrument for approximating CEO empathy.

“While empathy has historically been tough to measure at scale, our framework captures fine-grained behavioral cues—resembling emotional mimicry—and validates them by way of real-world settings,” mentioned Cui. “Using this measure, we discover that extra empathetic CEOs are related to extra prosocial company insurance policies—for instance, decrease office harm charges and extra equitable compensation insurance policies.

“Moreover, CEO empathy seems to contribute to simpler disaster administration, which in flip might assist improve agency worth. While preliminary, these findings mark to the potential organizational significance of empathy as a management trait. We see this as a step towards increasing how gentle expertise may be studied in real-world settings.”

The researchers hope that the framework they developed will additional advance accessible FinTech and psychology analysis instruments designed to review folks’s traits, communication types and habits, each in organizational and on a regular basis settings. In the longer term, it is also used to additional broaden the capabilities AI analytics platforms, permitting them to foretell folks’s empathy primarily based on their physique language.

“Our present study is an early step in exploring how video analytics might help quantify gentle traits like empathy, and there is definitely room to construct additional,” mentioned Cui. “One route we hope to pursue is making use of the framework to a broader set of video contexts—resembling social media clips, casual government communications, or cross-cultural settings—to enhance generalizability and perceive how empathy manifests in numerous environments.”

As a part of their future research, the researchers additionally plan to refine the framework they created, with the purpose of decreasing any potential biases in its interpretation of nonverbal cues. Finally, they might additionally attempt to devise comparable frameworks that may assess different human traits, talents or traits, resembling assertiveness, emotional stability and trustworthiness.

© 2025