Research
Consumer AI as Health Plan Decision Support
A benchmark of the July 2026 consumer AI lineup from OpenAI and Anthropic on a realistic health plan selection task, scored against a claims-based pricing engine. Measures what objective the models optimize, how accurate their cost estimates are, and how recommendations improve as consumers disclose more — or as structured insurance data is supplied.
Explore the findings →In progress
Can we design a decision support algorithm that improves consumer welfare and minimizes selection?
What happens to the risk pool if everyone gets good advice? A claims-based simulation quantifies the trade-off between consumer welfare and insurance market selection under different decision-support designs. — including a reimplementation of the HealthCare.gov cost estimator as a benchmark.
Get notified when it publishes →