Research / 2025–2026
Studies of AI behavior, human–AI collaboration, and the social contexts that shape both.
Multilingual LLMs / Sociocultural evaluation
A human-validated audit of 12 LLMs across English, French, and Chinese, separating surface cues from meaningful sociocultural patterns.
Human–AI collaboration / Creativity
An experiment with 700 participants reveals a trade-off between idea quality and satisfaction, and a tendency for human and AI evaluators to favor ideas from their own type.
Human–AI interaction / Prompting
Studying how prompt-engineering education affects journalists’ use of generative AI.
Human-centered AI / Recommendations
A user study of LLM-generated explanations that connect recommendations to their context.
LLM behavior / Decision-making
Examining how noise, adaptation, and strategy shape the fidelity of LLM-based decision-makers.
* Equal contribution. Preprints are labelled separately from conference publications.