Dan Jurafsky

Dr. Dan Jurafsky is a Professor at the Department of Linguistics within Stanford University's School of Humanities and Sciences. His work spans computational linguistics, natural language processing, and AI ethics, with a particular focus on language model behavior, speech recognition, and ethical implications of anthropomorphism in AI systems.

His recent research explores HumT for measuring human-like tone in LLMs, AnthroScore for anthropomorphism detection, and methods for improving low-resource language support through data augmentation and multilingual representation learning. He has also developed open-source tools like string2string for string algorithms.

Jurafsky's publications address critical issues in NLP, including grounding gaps in conversational models, causal interpretability in linguistic tasks, and representational biases in multilingual models. His work emphasizes interdisciplinary applications, from educational NLP tools to Sumerian transliteration datasets, while advocating for rigorous statistical power analysis and ethical model evaluation.