Mentor · Summer 2027
Muhua Huang
PhD Student, Stanford Graduate School of Business
About
Muhua is a PhD student studying Computational Social Science at Stanford University and a member of the Stanford-Berkeley Computational Culture Lab. Muhua’s research treats AI systems as active social participants, studying culture, values, cognition, and individual differences in AI using psychometric and computational social science methods. Muhua has previously conducted research on AI evaluation as a Student Researcher at Google DeepMind and on AI alignment at Microsoft Research Asia.
Research directions
- Evaluating the social and psychological properties of AI systems (e.g., personality, cognition, social behavior)
- Value and safety evaluation, and alignment of AI systems
Example project
Illustrative. Final projects are confirmed before applications open.
Psychometric validation of AI ‘personality’: a fellow would administer established personality or value inventories to several LLMs under varied prompts and personas, then test whether the responses show the same factor structure, reliability, and measurement invariance seen in human samples. The second stage would test predictive validity, asking whether trait scores predict how the same models behave as agents in downstream tasks such as negotiation or workplace simulations.
Methods they can supervise
- Psychometrics and measurement (scale development, factor analysis, SEM)
- LLM evaluation and benchmarking
- Generative agent-based simulations
- Computational text analysis and embeddings
- Survey experiments
Who they’re looking for
Fellows with an academic research background in psychology, sociology, cognitive science, or computer science, who have interdisciplinary interests and a clear sense of what they want to pursue.
Mentoring style
Remote. A weekly 30–60 minute meeting plus async communication on Slack.
Profile last updated:



