MIT researchers have developed a generative AI (GenAI) tool for databases that lets users analyze data and make predictions about future data or fill in missing information. The GenAI tool has been dubbed GenSQL, and it is an extension of the SQL programming language that integrates probabilistic programming with traditional database searches.

In addition, the tool lets business users ask complex questions and then combines actual data with probabilistic reasoning, generating more nuanced insights about a product or service. Developers can leverage GenSQL to employ probabilistic modeling in databases without expertise in probabilistic programming.

The MIT team created GenSQL with the goal of simplifying, managing, and analyzing data from various sources. The researchers stated that existing probabilistic programming systems fail to support complex database queries and do not combine tabular data with generative models effectively.

With GenSQL, users can upload their data and models to GenSQL, which automatically integrates them. Users can also attempt task searches that include data cleaning and synthetic data generation and develop custom models for harmonization across different data sources.

On evaluation, researchers found the tool to be more concise and less error-prone while detecting database anomalies compared to prior systems, and it conducts tasks seven times faster.