About the Project Pareto.AI is a human data-collection platform connecting leading AI researchers with trusted industry experts to collaborate on AI alignment, safety, and training projects. We are partnering with a frontier AI lab to evaluate an AI model's ability to replicate empirical economics research findings.
Identify suitable causal economics papers with publicly available replication dataWrite prompts asking the AI model to replicate findings given a research question, dataset, codebook, and contextWrite rubrics to evaluate the AI model's performance across each step of the empirical pipeline:Data cleaningPhD in Economics (required)Familiarity with replication-friendly microdata — NLSY, ACS, CPS, administrative dataProficient in STATA, R, or PythonStrong understanding of empirical research workflow from raw data to published resultsBonus: experience with AI/ML tools or interest in AI evaluation
Active or former academic economist at a research universityFields: labor, health, development, public, environmental economics
Contribute to cutting-edge AI safety and alignment researchFlexible part-time remote work — task-based engagementCollaborate with a global network of economists and AI researchersCompetitive compensation per completed taskCompensation - $100/hr USD
If your background is a strong fit, you'll receive a direct link to join the project and complete your application.