How AI and Neuroscience Are Decoding the Art of Hip-Hop Dance

Where AI Meets Hip-Hop

Artificial Intelligence

8 MIN READ

Assistant Professor of Philosophy Ben Baker at Colby College is combining AI, neuroscience, and dance to explore human movement and cognition. Working alongside students Kayla Leonard ’25 and Hannah Junn ’25 at the Gordon Center for Creative and Performing Arts, Baker uses advanced equipment to capture dancers’ motions and analyze them with artificial intelligence.

During a recent session, dancers performed to Future’s song “Mask Off” while cameras recorded every move. The footage transforms into 3D models highlighting joints and limbs, which AI then examines. Baker focuses on two key questions: How do movements differ across dance genres like house and breakdance? And what changes occur in a dancer’s body mechanics as they learn new moves or practice over time?

Baker, who is also a dancer, tests these ideas firsthand. When working with student dancers, he quickly picks up choreography and joins in, demonstrating a hands-on approach to his research.

Remixing Philosophy, Science, and Art

Hannah Junn ’25, a Korean-American dancer and co-founder of the hip-hop group Cozy at Colby, finds a unique blend of culture, dance, and technology in this research. Cozy meets weekly to teach hip-hop moves, blending fitness and fun. Junn was intrigued by the intersection of philosophy, AI, and dance, initially wondering how these fields could merge.

Baker’s path made the connection clearer. With a background in philosophy and cognitive science, he found dance offered a practical lens for studying embodied cognition—how intelligence expresses itself through physical movement. His experience dancing with the Academy of Phresh in Philadelphia added real-world context to his academic work.

He explains that AI serves as a powerful statistical tool to analyze movement patterns, distinct from human intelligence. While AI can identify patterns, it lacks the human capacity for creativity and experience. Recognizing this distinction helps preserve the value of human intelligence while leveraging AI’s analytical strengths.

Teaching AI Hip-Hop Styles

Baker joined Colby’s faculty in 2023 and quickly connected with computer science major and dancer Beste Kuruefe ’26. Kuruefe, passionate about patterns and rules, embraced the opportunity to apply AI to dance research. Together, they used an existing dataset of 1,408 3D dance movement models covering 10 different hip-hop styles.

The goal was to train a machine learning model to classify dance genres. Rather than letting the AI develop its own features—which often results in an opaque “black box” model—they selected 17 interpretable features based on movement knowledge. These included joint motions like ankle and wrist movement, and body twists.

This approach allowed them to understand which features were most important for distinguishing styles. Expandability—the distance of joints from the sacrum—emerged as the key feature across all styles. Mapping these styles in a 3D feature space revealed clusters of similar genres and an intriguing empty space that could hint at an undiscovered or non-preferred movement style.

Next Steps

With support from Cozy at Colby, Baker plans to deepen this research by capturing new data and observing dancers as they learn and refine moves over time. This longitudinal approach aims to identify how movement features evolve during skill acquisition, potentially informing new dance education and training tools.

Real-time AI analysis is another goal. Such a system would provide immediate feedback on detected movement features, aiding choreographers and dancers in exploring the feature space more creatively. This could help artists break out of repetitive patterns and develop innovative sequences.

Kuruefe is eager to contribute to these developments. Baker notes that tools like this could inspire new directions in choreography by highlighting unexplored or underutilized movement styles.

Junn appreciates the novel blend of dance and technology, remarking on the uniqueness of applying research to something as dynamic and expressive as hip-hop dance.

For those interested in the intersection of AI and the arts, this research illustrates practical applications of machine learning beyond traditional domains, offering new ways to analyze and enhance human creativity.