Nathanael J

I study when people delegate their cognitive burden to generative AI, how that affects their agency, and how incentives shape that balance. But we also constantly interact with predictive algorithms too! I’m interested in similar questions around incentive design for algorithms at large.

• Alignment has a Fantasia Problem

  Working paper, 2026

• Incentives shape how humans co-create with generative AI

  Working paper, 2026

An AI system's value lies not only in its ability to produce a correct output, but also in how it shapes human thinking and decision-making. I study these questions from both a normative perspective (what should count as good performance?) and a statistical one (what can we actually infer from the evaluations we run?).

• The Subjectivity of Monoculture

  Working paper, 2026

• Position: AI Evaluations Should be Grounded on a Theory of Capability

  ICML (Position), 2026

I'm interested in developing decision making tools that align with stakeholder priorities—particularly for interpretability and fairness. I also use machine learning and large-scale data to generate policy insights.

• Learning Optimal Prescriptive Trees from Observational Data

  Management Science, 2026

• Not (Officially) in My Backyard: Characterizing Informal Accessory Dwelling Units and Informing Housing Policy with Remote Sensing

  JAPA, 2024