Neel Guha

I am a fourth year JD-PhD student in Computer Science at Stanford University (advised by Chris Ré) and a Graduate Student Fellow at the RegLab. I graduated with a MS in Machine Learning from Carnegie Mellon University (‘19) and a BS (with Honors) in Computer Science from Stanford University (‘18). I am grateful to be supported by the Stanford Interdisciplinary Graduate Fellowship (SIGF) and the HAI Graduate Fellowship.

My computer science research explores (1) legal applications of large language models, and (2) methods for enabling machine learning systems to better leverage structure (e.g., type information, knowledge base triples, embeddings). My recent work includes leading the creation of LegalBench, a large scale open effort to develop benchmark tasks for evaluating legal reasoning in LLMs.

My legal research agenda has two focuses. First, I am interested in how we can use machine learning tools to push the boundaries of empirical legal studies. Traditionally, the scope of questions empirical legal scholars have been able to study has been limited by what types of datasets can be feasibly developed by hand. Machine learning offers an exciting path towards automatically, cheaply, and robustly constructing large scale legal datasets, thus allowing us to explore substantively broader questions (e.g., how many private rights of action exist across all state codes?).

Second, I am interested in the interplay between regulatory policy, tort liability, and AI governance. As AI proliferates across sectors (e.g., healthcare, finance, energy), the types of AI systems deployed will be guided by standards fashioned by both regulators and courts. My work attempts to describe how these stakeholders can approach governance choices in a technically and legally principled way. For instance, I’ve written about how courts might assess liability for medical AI, and the technical/institutional tradeoffs raised by different regulatory mechanisms.