Can artificial intelligence help identify the unconscious biases that shape decisions in high-stakes moments? Researchers at Fort Lewis College think so, and they're using AI tools to examine how racial bias may influence referee behavior in the Ultimate Fighting Championship (UFC).
With funding support from the AI Institute’s Elevate AI incubator program, the Social Perception Experimentation (SPEx) Lab uses machine learning tools to streamline data collection and deepen its analyses of human behavior.
Alex Borgella, Ph.D., associate professor of Psychology and student researchers in the SPEx Lab, are analyzing thousands of UFC bouts to uncover patterns in how and when referees stop fights. Their early findings suggest referees may take longer to intervene after Black, Latine, and other athletes of color are knocked unconscious or rendered defenseless—raising questions about athlete safety and equity.
“It’s a really nice, controlled way to look at these biases because the decisions have to be made really quickly,” Borgella said. “You can’t really say that the referee’s got something out for the fighter due to explicit racial animus—it’s literally their job to make that decision in the blink of an eye, which makes it a great window into implicit bias.”
The project uses machine learning tools to streamline data collection and automate complex coding tasks. Before integrating AI, compiling and analyzing UFC data by hand took over a year. Now, the team can complete the same work in a matter of weeks.
Research assistant Rowan Freeman, a third-year Psychology major at FLC, led the development of a Python script that scraped thousands of fights and used AI-powered facial recognition to identify the perceived racial affiliation of fighters and referees. Though initially unsure about relying on AI, Freeman said the experience gave her valuable hands-on research skills.
“I didn’t want [AI] to take the place of someone learning or using their skills,” she said. “But I still got experience. It was surprisingly hands-on. I think it made the research more accessible.”
Borgella praised Freeman’s work noting that the programming language she is using (Python) is notoriously difficult for beginners: “It took her about a week to figure it out [the programming she created to scrape UFC data], which is crazy considering she had no prior experience with Python whatsoever.” he said
The team plans to expand their dataset with new bouts and additional variables such as tattoo coverage and betting odds to examine how other cues may influence referee behavior. The research combines psychological theory, real-world data, and cutting-edge technology to explore the dynamics of race, perception, and decision-making.
“This funding with AI in mind specifically is really going to be the game changer,” Borgella said. “It’s transforming the way we do our work.”
The SPEx Lab’s work supports FLC’s mission to advance equity, innovation, and experiential learning through faculty-led research that places students at the center of inquiry and discovery.