OpenAI’s Economic Research Team just dropped GABRIEL, an open-source toolkit that uses GPT to convert messy, qualitative data—like interviews, syllabi, or social media posts—into quantitative scores. Researchers can ask questions in plain English (e.g., ‘How family-friendly is this job listing?’) and get consistent, numerical answers across millions of documents. This saves weeks of manual labeling, letting experts focus on analysis instead of drudgery. The tool also handles dataset merging, deduplication, and privacy-preserving deidentification. In benchmarks, GPT proved highly accurate for these labeling tasks. Why it matters: By slashing the time and cost of analyzing qualitative data, GABRIEL could unlock entire new categories of social science research that were previously impossible due to scale, giving us richer insights into human behavior.