Harvard, UCLA Researchers Use AI to Analyze Passion in Public Speaking

Sesh describes itself as a “neuroscience-based people development company.” Its cornerstone product is a tool called Empath, which uses AI to analyze video content, delivering insights on behavior and communication patterns. Now, Sesh is announcing a new collaborative project with the Harvard Business School and the UCLA Anderson School of Management that is using the firm’s advanced AI video analysis tools to measure how the passion of public speakers affects perception, attitudes and behaviors.

Sesh’s Empath tool uses AI to analyze emotions and behaviors in video content.

Sesh was founded in 2019 with a focus on training and development sessions with an eye toward using video data from those sessions to train an AI. When the pandemic struck, those plans accelerated, giving rise to Empath – which released in July 2020 – and, by August, netting $1.6 million in pre-seed funding for the company. “With most of the world now communicating by video, being able to truly understand the non-verbal cues, situational context and culture of your meeting participants is crucial to quickly establishing trust, and building meaningful relationships,” co-founder and CEO David Dorfman said on the heels of that announcement.

That data and the ensuing insights, of course, are not only valuable to enterprise customers. Joyce He, an incoming assistant professor at UCLA’s Anderson School of Management, and Jon M. Jachimowicz, an assistant professor of organizational behavior at Harvard Business School, also saw the potential.

The project will use Sesh’s AI and proprietary AI qualitative coding tools, along with the company’s deep data analytics and video, audio, language, psychographic, and cultural datasets. The research duo, equipped with their own 1.8 million minutes of video, will use these tools to test how passion intersects and interacts with the reactions to and outcomes of public speaking. Specific points of interest for the researchers include identifying which of public speakers’ behaviors are associated with people’s perception of passion; identify different varieties of passion and their relative outcomes; and determine whether passion has different effects along demographic lines like gender, race, and socioeconomic status.

“Beyond their theoretical relevance, understanding whether and when displays of passion predict outcomes also has important practical implications for leadership and speech training. For instance, after we identify which behaviors are most associated with perceptions of passion, leaders could learn these to facilitate their expressions of passion and better inspire their people. Our collaboration with Sesh will allow us to extend the limits of this (artificial intelligence) technology to more complex emotions,” Jachimowicz said.