The National Artificial Intelligence (AI) Research Resource (NAIRR) Task Force convened its third virtual, public meeting on October 25 to further develop a vision and implementation plan for the NAIRR—a national cyberinfrastructure that would democratize access to resources and tools that fuel AI research and development (R&D), making AI R&D equitable and accessible to more American researchers, sparking innovation and economic development, and helping build an AI workforce for the future. Through a series of meetings, the Task Force is working toward consensus recommendations on a NAIRR implementation plan and roadmap that will be provided to Congress in 2022.
Co-Chairs Dr. Lynne Parker, Director of the National AI Initiative at the White House Office of Science and Technology Policy (OSTP), and Dr. Erwin Gianchandani, Senior Advisor for Translation, Innovation, and Partnerships at the National Science Foundation (NSF), opened the meeting by announcing the posting on AI.gov of responses received to the RFI on an Implementation Plan for a National AI Research Resource. More than 80 unique responses were received by the October 1 deadline from a wide range of academic, private sector, non-profit, and governmental stakeholders. This valuable input will help inform the Task Force’s work as it develops a NAIRR implementation plan.
The first session of the meeting engaged Task Force members in discussions of draft recommendations for the computational resources that should be available through the NAIRR, as well as the appropriate governance and administration of the NAIIR. These discussions built upon ideas provided by a range of external experts and practitioners during the first and second Task Force meetings on July 29 and August 30, respectively. The Task Force agreed that the NAIRR should federate computational resources, embodying a mix of cloud and on-premise resources, experimental and production environments, and core and edge computing. The Task Force also agreed on scaling the resources in two ways: the NAIRR should address the need both for high-end computational capabilities for large-scale AI problems, as well as widely-distributed resources to provide computational capacity for many users simultaneously.