Daniel Low

Daniel Low is a Research Scientist and tenure-track Principal Investigator at the Child Mind Institute in New York City and a Visiting Scholar in the Department of Psychology at Harvard University.

He directs a lab at the Child Mind Institute that leverages LLMs and causal inference to analyze text and speech data for different mental health applications: detecting suicide risk from journaling and social media posts, uncovering how insight emerges throughout clinical interviews after meditation and psychedelic sessions, and evaluating human preferences and outcomes for different AI chatbot responses when used for psychological support. His work seeks to inform causation questions in legal cases and policy around AI and social media. 

He recently completed a NIMH T32 Postdoctoral Fellowship on causal inference for suicide prevention mentored by Matthew Nock (Department of Psychology, Harvard University) and Miguel Hernán (Department of Epidemiology, Harvard Chan School of Public Health) at the CAUSALab. He received his PhD in Speech and Hearing Bioscience and Technology from Harvard University in 2024 under the supervision of Satrajit Ghosh (MIT). His dissertation focused on detecting dozens of suicide risk factors in text and speech using LLMs, speech processing, and machine learning. His prior training was in Argentina, Italy, and the Netherlands in cognitive science and natural language processing.

He co-founded the Harvard-MIT Speech and Language Biomarker Interest Group which organizes talks and discussions in this field, and he has taught machine learning and natural language processing at Harvard and other universities.

He aspires to leverage AI, causal inference, and phenomenology to reduce suffering and enhance flourishing in a scalable and equitable way.