Inquiry and innovation are rarely solitary undertakings. In societies as diverse, interconnected, and AI-mediated as ours, how we understand and shape the world depends on our social positions, institutional norms, and technological infrastructures. Most of my work asks how these social and technological factors shape what we come to know and do, and how their influence should be evaluated and governed to serve epistemic and ethical ends. I pursue this across three connected lines of research, combining philosophical analysis with computational and empirical methods.
Sociotechnical Epistemology of AI
AI systems are transforming how knowledge is sought, produced, circulated, and debated. How to anticipate and mitigate the risks these transformations pose to validity, diversity of thought, and epistemic justice, and instead make our knowledge ecosystem more resilient? I work on these questions at three scales. At the level of individual cognition, I approach AI as an epistemic affordance (Philosophical Issues 2026), asking how it can and should restructure opportunities for epistemic agency. At the level of institutional practices, I ask what follows when systems are used in novel epistemic roles, such as adjudicating contested meaning (FAccT 2026) or simulating human participants (PNAS 2025), and I develop frameworks for evaluating those uses (AAAI 2026). At the level of knowledge communities, I study how AI’s emergent effects on collective inquiry can be understood and steered (Philosophy of Science 2026).
Normative Foundations of Responsible AI
AI systems are informing consequential institutional decisions about how benefits and burdens are allocated. How to ensure these applications respect epistemic values like validity and reliability, and ethical ones like justice and fairness? I work on characterizing the kinds of bias that threaten these values and mapping their sources across the AI lifecycle (Philosophy Compass 2021). Judgments about whether such biases are responsibly managed are often framed in terms of fairness or performance metrics defined over model properties. I examine how this restricted framing leads to misguided responses to AI risks (AIES 2020, AIES 2021, CJP 2022), and develop procedural and system-level frameworks that support implementing our values without collapsing them into distorted proxies (Big Data & Society 2022, FAccT 2022, BJPS 2025). I also carry this work into practice, contributing to the NIST AI Risk Management Framework.
Diversity and the Social Dimensions of Inquiry
Inquiry is largely carried out in social settings, by groups like design teams, deliberative mini-publics, and scientific communities. How to organize these groups and their social environments to make productive use of diverse perspectives and capabilities? I examine how contextually relevant senses of diversity should be conceptualized and measured (European Journal for Philosophy of Science 2018), as well as the interpersonal (Philosophy of Science 2022, Synthese 2021), structural (CogSci 2022), and technological (Synthese 2025) conditions that shape the epistemic and ethical dynamics of diverse groups. I bring this work to bear on technology design and social policy, including how attention to diversity and disagreement improves integrity of AI systems and ecosystems (Big Data & Society 2022, FAccT 2025).
Peer-reviewed Publications
See my Google Scholar page for a full list as well as recent manuscripts and preprints.
Aspirational Affordances of AI
Philosophical Issues, Accepted
with: M. Magnani
Navigating Epistemic Monocultures in AI-Driven Science: A Simulation Study
Philosophy of Science, 2026
with: J. O’Brien, H. Rubin
Ambiguity Collapse by LLMs: A Taxonomy of Epistemic Risks
ACM Conference on Fairness, Accountability, and Transparency (FAccT), 2026
with: S. Gur-Arieh, A. Wang
Should You Use LLMs to Simulate Opinions? Quality Checks for Early-Stage Deliberation
AAAI Conference on Artificial Intelligence, 2026
with: T. Neumann, M. De-Arteaga
Disciplining Deliberation: A Sociotechnical Perspective on Machine Learning Trade-Offs
The British Journal for the Philosophy of Science, 2025
Authenticity and Exclusion: A Simulation Study of How Social Media Algorithms Shape Visibility in Epistemic Communities
Synthese, 2025
with: N. J. Akpinar
Take Caution in Using LLMs as Human Surrogates
Proceedings of the National Academy of Sciences, 2025
with: Y. Gao, D. Lee, G. Burtch
The Value of Disagreement in AI Design, Evaluation, and Alignment
ACM Conference on Fairness, Accountability, and Transparency (FAccT), 2025
with: W. Fleisher
Diversity in Sociotechnical Machine Learning Systems
Big Data & Society, 2022
with: M. De-Arteaga
Algorithmic Fairness and the Situated Dynamics of Justice
Canadian Journal of Philosophy, 2022
with: Z. C. Lipton, D. Danks
Diversity, Trust, and Conformity: A Simulation Study
Philosophy of Science, 2022
with: D. Steel
Justice in Misinformation Detection Systems: An Analysis of Algorithms, Stakeholders, and Potential Harms
ACM Conference on Fairness, Accountability, and Transparency (FAccT), 2022
with: T. Neumann, M. De-Arteaga
Diversity and Homophily in Social Networks
Proceedings of the Annual Conference of the Cognitive Science Society (CogSci), 2022
with: H. Rubin
Algorithmic Bias: Senses, Sources, Solutions
Philosophy Compass, 2021
with: D. Danks
Norms in Counterfactual Selection
Philosophy and Phenomenological Research, 2021
Fair Machine Learning Under Partial Compliance
AAAI/ACM Conference on AI, Ethics, and Society (AIES), 2021
with: J. Dai, Z. C. Lipton
Information Elaboration and Epistemic Effects of Diversity
Synthese, 2021
with: D. Steel, B. Crewe, K. Gillette
The Many Faces of Attention: Why Precision Optimization Is Not Attention
In The Philosophy and Science of Predictive Processing, Bloomsbury, 2021
with: M. Ransom
Affect-Biased Attention and Predictive Processing
Cognition, 2020
with: M. Ransom, J. Markovic, J. Kryklywy, E. Thompson, R. M. Todd
Algorithmic Fairness from a Non-Ideal Perspective
AAAI/ACM Conference on AI, Ethics, and Society (AIES), 2020
with: Z. C. Lipton
Multiple Diversity Concepts and Their Ethical-Epistemic Implications
European Journal for Philosophy of Science, 2018
with: D. Steel, K. Gillette, B. Crewe, M. Burgess
Attention in the Predictive Mind
Consciousness and Cognition, 2017
with: M. Ransom, C. Mole
The Kantian Brain: Brain Dynamics from a Neurophenomenological Perspective
Current Opinion in Neurobiology, 2015
with: E. Thompson
Policy work
Red Teaming AI: The Devil Is in the Details
Tech Policy Press, 2024
with: D. Hadfield-Menell, L. Belli
Artificial Intelligence Risk Management Framework (AI RMF 1.0)
National Institute of Standards and Technology, 2023
contributor
Pathways to Digital Justice
World Economic Forum white paper, 2021
lead co-author
Sina Fazelpour 