Unlocking Dataset Value for AI-Enabled Scientific Discovery (AI Datasets)

The National Science Foundation (NSF) has launched the "Unlocking Dataset Value for AI-Enabled Scientific Discovery (AI Datasets)" program to maximize the scientific utility of existing datasets using artificial intelligence. This joint directorate initiative aims to apply advanced AI techniques to feature extraction, metadata generation, and multi-dataset integration. By building automated, robust data pipelines and harmonizing data structures, the program seeks to facilitate interdisciplinary inquiries and unlock novel insights far beyond original data collection parameters. The solicitation supports three main award categories: Flagship awards up to $5,000,000, Impact awards up to $2,000,000, and Planning grants up to $200,000, drawing from an overall anticipated budget of $60,000,000 to $100,000,000. Awardees must address dataset governance, integrity, and security while leveraging national infrastructure such as the National AI Research Resource (NAIRR). Research proposing de novo data collection is out of scope for this program.