Publications

Our research focuses on stereotype bias in humans, using four methodological approaches. You can browse by keyword: data-driven discovery, model-based mechanisms, large-scale behavioral experiments, and bias in AI.

Explicitly unbiased large language models still form biased associations.
Xuechunzi Bai, Angelina Wang, Ilia Sucholutsky, Tom Griffiths
PNAS, 2025 [pdf] [project] [slides]

Globally inaccurate stereotypes can result from locally adaptive exploration.
Xuechunzi Bai, Susan Fiske, Tom Griffiths
Psych Science, 2022 [pdf] [project] [slides]

As diversity increases, people paradoxically perceive social groups as more similar.
Xuechunzi Bai, Miguel Ramos, Susan Fiske
PNAS, 2020 [pdf] [project] [slides]

Collective bias emerges even from rational social learning
Bufan Gao, Xuechunzi Bai
arXiv, Under review [pdf]

Large language models develop novel social biases through adaptive exploration
Addison Wu, Ryan Liu, Xuechunzi Bai, Tom Griffiths
arXiv, Under review [pdf]

Aligned but blind: Alignment increases implicit bias by reducing awareness of race.
Lihao Sun, Chengzhi Mao, Valentin Hofmann, Xuechunzi Bai
ACL Main, 2025 [pdf] [project]

Measuring machine learning harms from stereotypes requires understanding who is being harmed by which errors in what ways.
Angelina Wang, Xuechunzi Bai, Solon Barocos, Su Lin Blodgett
FAccT, 2025 [pdf]

Explicitly unbiased large language models still form biased associations.
Xuechunzi Bai, Angelina Wang, Ilia Sucholutsky, Tom Griffiths
PNAS, 2025 [pdf] [project] [slides]

Learning too much from too little: False face stereotypes emerge from a few exemplars and persist via insufficient sampling.
Xuechunzi Bai, Stefan Uddenberg, Brandon Labbree, Alex Todorov
JPSP, 2025 [pdf]

Costly exploration produces stereotypes with dimensions of warmth and competence.
Xuechunzi Bai, Tom Griffiths, Susan Fiske
JEP:G, 2025 [pdf] [model] [slides]

The limitations of machine learning models for predicting scientific replicability.
Molly Crockett, Xuechunzi Bai, Sayash Kapoor, Lisa Messeri, Arvind Narayanan
PNAS Letters, 2023 [pdf]

Humans perceive warmth and competence in Artificial Intelligence.
Kevin McKee, Xuechunzi Bai, Susan Fiske
iScience, 2023 [pdf]

Globally inaccurate stereotypes can result from locally adaptive exploration.
Xuechunzi Bai, Susan Fiske, Tom Griffiths
Psych Science, 2022 [pdf] [project] [slides]

A spontaneous stereotype content model: Taxonomy, properties, processes, and prediction.
Gandalf Nicolas, Xuechunzi Bai, Susan Fiske
JPSP, 2022 [pdf]

Warmth and competence in human-agent cooperation.
Kevin McKee, Xuechunzi Bai, Susan Fiske
AAMAS, 2022 [pdf]

Urban space and social cognition: The effect of urban space on intergroup perceptions.
Kim Knipprath, Maurice Crul, Ismintha Waldring, Xuechunzi Bai
ANNALS, 2021 [pdf]

Cosmopolitan morality tradesoff ingroup for the world, seperating benefits and protection.
Xuechunzi Bai, Varun Gauri, Susan Fiske
PNAS, 2021 [pdf]

As diversity increases, people paradoxically perceive social groups as more similar.
Xuechunzi Bai, Miguel Ramos, Susan Fiske
PNAS, 2020 [pdf] [project] [slides]

Comprehensive stereotype content dictionaries using a semi-automated method.
Gandalf Nicolas, Xuechunzi Bai, Susan Fiske
EJSP, 2020 [pdf]

Vertical and horizontal inequality are status and power differences.
Susan Fiske, Xuechunzi Bai
Current Opinion, 2020 [pdf]

Stereotypes as historical accidents: Images of social class in post-communist versus capitalist societies.
Lucy Grigoryan, Xuechunzi Bai, Federica Durante, Susan Fiske, Marharyta Fabrykant, Anna Hakobjanyan, Nino Javakhishvili, Kamoliddin Kadirov, Marina Kotova, Ana Makashvili, Edona Maloku, Olga Morozova-Larina, Nozima Mullabaeva, Adil Samekin, Volha Verbilovich, Illia Yahiiaiev
PSPB, 2020 [pdf]

Exploring research methods blogs in psychology: Who posts what about whom.
Gandalf Nicolas, Xuechunzi Bai, Susan Fiske
PPS, 2019 [pdf]

Admired rich or resented rich? How two cultures vary in envy.
Sherry Wu, Xuechunzi Bai, Susan Fiske
JCCP, 2018 [pdf]