A new initiative called Project ProtoStar will help universities and research organizations publish publicly funded datasets, models and code in formats that AI systems and researchers can reuse. The first open challenge, hosted by Carnegie Mellon University, opens on November 2.
Founding participants include Carnegie Mellon, Purdue, Georgia Tech, Arizona State and George Washington University through the Open Forum for AI. The Astera Institute, Hugging Face and the Open Research Community Accelerator are also involved. Hugging Face will provide infrastructure, while participating organizations sign agreements supporting release of suitable scientific artifacts on open systems.
The project addresses a gap between public access and practical usability. Research data may technically be available while remaining scattered across repositories, stored in incompatible or legacy formats, or missing documentation needed to interpret it. ProtoStar plans to promote machine-actionable packaging and demonstrate its value through open challenges.
The initiative does not automatically resolve licensing, privacy, consent or data-quality problems. Each institution still has to decide what can be shared and document how it was produced. Its near-term test will be whether the challenges yield durable datasets and code that outside teams can reproduce, rather than another discovery portal pointing to difficult-to-use files.