OpenAI has launched GABRIEL, an open-source toolkit designed to help researchers transform qualitative data into quantitative measurements at scale. The tool, developed by OpenAI's Economic Research Team, uses GPT to analyze unstructured text and images, making it easier for economists, social scientists, and data scientists to study qualitative data systematically.

Qualitative data—from interviews and syllabi to social media and photographs—contains rich insights about human experiences but has traditionally been difficult to analyze rigorously. GABRIEL addresses this challenge by allowing researchers to describe measurement goals in everyday language, such as "how family-friendly is this job listing?" The tool then applies these criteria consistently across thousands or millions of documents, returning standardized scores. This automation reduces the time spent on repetitive data labeling while preserving the expertise needed for choosing what to measure and validating results.

Beyond measurement capabilities, GABRIEL includes practical tools for merging mismatched datasets, deduplication, passage coding, theory development, and deidentifying personal information. The Python library is available now with a tutorial notebook, requiring minimal technical background. OpenAI plans to continue improving GABRIEL based on academic feedback, aiming to help more researchers incorporate the richness of qualitative data and human stories into their work.