Making education data accessible

OpenAI Blog Products

Summary

Zelma, a GPT-4 powered research assistant developed by Dr. Emily Oster and her team at Brown University in partnership with Novy, makes standardized test data accessible to parents, teachers, administrators, and policymakers by allowing natural language queries about student performance across districts and demographics.

Zelma uses GPT-4 to make education data accessible.
Original Article
View Cached Full Text

Cached at: 04/20/26, 02:44 PM

# Making education data accessible Source: [https://openai.com/index/zelma/](https://openai.com/index/zelma/) Zelma, a GPT‑4 powered research assistant, is making education data accessible to parents, teachers, school administrators, and policymakers across the United States\. The problem they’re solving? While standardized tests are issued to every student in grades 3–8 in the U\.S\., data about how students perform remains largely untapped and underutilized, scattered across different sources and formats\. Test data offers powerful insights into how well students are learning in English language arts \(ELA\) and math, across district and demographic group levels\. But crucial insights that could shape the future of education are effectively beyond the reach of most parents, journalists, educators, and policymakers\. Recognizing this gap, Dr\. Emily Oster, a Brown economist and author, set out to change the narrative\. “Zelma makes it possible for parents and policymakers alike to use plain language to get instant, tailored educational insights on what matters most to them,” Dr\. Oster explained\. How did Zelma accomplish this? Dr\. Oster’s team of student researchers at Brown University spent a year gathering the data and meticulously cleaning it into one uniform format\. Zelma then worked with[Novy⁠\(opens in a new window\)](https://www.novy.ai/)to bring the data to life using OpenAI’s API\. Novy leveraged[function calling⁠\(opens in a new window\)](https://platform.openai.com/docs/guides/function-calling)to tell GPT‑4 which visuals and fields to choose from in displaying the data\. For Zelma’s "Ask a Question" view, they[fine\-tuned⁠\(opens in a new window\)](https://platform.openai.com/docs/guides/fine-tuning)a model to create a data\-aware type\-ahead that suggests questions with available data\. And they[embedded⁠](https://openai.com/blog/introducing-text-and-code-embeddings)and stored known\-good example graphs in a vector database to improve accuracy on difficult edge cases\. The biggest challenge was trying to predict how people would phrase their questions, and keeping the prompts within Zelma’s scope of knowledge\. “We addressed this through intentional design choices in Zelma’s user experience that nudge people to ask questions Zelma can answer,” Dr\. Oster explained\. Key features include: - **Suggesting example prompts\.**Because Zelma covers only a specific domain of education data, these suggestions help users understand Zelma’s scope of knowledge and how to phrase their prompts\. - **Making all questions**[**public\.**⁠\(opens in a new window\)](https://www.zelma.ai/featured)This helps people learn from what others have asked while also preventing users from submitting high volumes of irrelevant queries\. - **Showing the SQL code\.**Zelma provides its logic to help you quickly verify its answers\. For non\-coders, the factors are also explained in plain language\. - **Explaining the context\.**Zelma gives context through definitions and flagging notable events—such as changes in the state assessment—to help you understand any external factors that might have influenced the data in a given year\.

Similar Articles

Using GPT-4 to improve teaching and learning in Brazil

OpenAI Blog

Arco Educação, Brazil's largest educational operating system, is partnering with OpenAI to launch the Teacher Assistant, an AI tool powered by GPT-4 that helps teachers create personalized lesson plans for students with diverse learning needs, significantly reducing time spent on administrative tasks and lesson planning.

Personalizing education with ChatGPT

OpenAI Blog

Arizona State University integrates ChatGPT Edu across teaching, research, and operations, launching an AI Innovation Challenge that received over 400 proposals from faculty and staff to enhance personalized learning and student outcomes.

Scaling social science research

OpenAI Blog

OpenAI releases GABRIEL, an open-source toolkit that uses GPT to convert unstructured qualitative data (text, images) into quantitative measurements for social scientists and economists. The tool enables researchers to analyze large-scale qualitative datasets more efficiently by automating repetitive labeling tasks while preserving the richness of human data.

Powering virtual education for the classroom

OpenAI Blog

Khan Academy is partnering with OpenAI to integrate GPT-4 into its educational platform to enhance student learning through deeper conceptual understanding and personalized tutoring, while also helping teachers create customized instructional materials and track student progress.

Empowering teams to unlock insights faster at OpenAI

OpenAI Blog

OpenAI has developed an internal research assistant that combines dashboards with a conversational GPT-5 interface to help teams analyze millions of support tickets and generate insights in minutes instead of weeks. The tool democratizes data analysis across teams, allowing non-technical users to ask questions in plain language and get actionable reports on product feedback, customer sentiment, and trends.