Research from the MIT Media Lab finds that human-composed music more effectively elicits target emotions, even when listeners say they prefer AI.

As generative AI music becomes increasingly sophisticated and commercially attractive, new research from the MIT Media Lab's Opera of the Future group suggests the field may be measuring its impact the wrong way.

The paper, Understanding Listener Perceptions of AI and Human-Composed Music in Emotional Applications, was authored by Kimaya Lecamwasam, a researcher in the group, and Tishya Ray Chaudhuri of Myndstream, a global health and wellness music company. It will be presented at NIME 2026 (New Interfaces for Musical Expression) in London and further results will be presented at SMPC 2026 (Society for Music Perception and Cognition) in Chicago.

In a controlled experiment, 152 participants listened to instrumental music designed for either calming or energizing wellness contexts. Half the tracks were composed by Myndstream's in-house composers; half were generated by SunoAI. Some participants were correctly told each track's origin, others were intentionally given incorrect labels, and others heard the music without any labels at all — allowing the researchers to examine how authorship, and belief about authorship, shape emotional response.

One of the clearest findings was a disconnect between preference and emotional efficacy: participants often said they preferred AI-generated music, yet human-composed pieces were more effective at eliciting the targeted emotional states. Participants were also frequently more emotionally moved by a track when they believed it was composed by a person — even when it wasn't.

"These findings point to a more nuanced emotional landscape that current generative music models don't fully capture," said Lecamwasam. 

We need more human-centric approaches — ones that measure not just what people say they like, but how music actually makes them feel, both psychologically and physiologically.

The research team plans to extend the work to explore genre-specific and cross-cultural responses, physiological indicators such as heart rate and skin conductance, and how expert musicians perceive AI-generated music differently from general audiences.

Reflecting on the findings, Jordan Galvan, Head of Music at Myndstream, said: 

This research reinforces what we see every day: music's true power lies in its ability to move people and resonate with them on a deep, emotional level. 

"As AI becomes more prominent, emotional connection will be the difference between what's heard and what's felt."

Myndstream supported the research through a gift to the Opera of the Future group at the MIT Media Lab. The company's contributions also extended into the design of the study itself: drawing on his industry expertise, Jordan Galvan crafted the creative briefs that guided both the AI generation and the human composition, and Myndstream's in-house composersd brought those briefs to life in the human-composed tracks that participants heard.