Safe-AI: sandbox for accelerated feedback and experimentation for AI innovation

The National Science Foundation awarded the University of Utah $2.164 million on July 1, 2026, under the NSF Technology, Innovation, and Partnerships program (CFDA 47.084) to develop and launch a two-year AI innovation testbed called SAFE-AI.

The University of Utah will coordinate the design and deployment of the SAFE-AI Sandbox, a testbed supported by a public-private partnership ecosystem that brings together state government, academia, and industry partners to advance AI research, adoption, and workforce development. The testbed will explore AI implementation across three use cases: AI-assisted radiological interpretation, mental health assessment and education, and runtime enforcement infrastructure. The work leverages AWS innovation sandboxes to provide secure, isolated, and auditable environments for experimentation and demonstration. The SAFE-AI Advisory Committee, comprising experts from Utah academic institutions, government entities, and industry partners including Mindshare Ventures, Ascend AI Labs, and Clarion AI Partners, will guide testbed development, deployment, and execution of the use cases and partnership models. The project will result in coordinated infrastructure, formalized public-private partnerships, and regulatory implementation strategies for industry adoption in complex, regulated sectors such as healthcare. The testbed will engage with the NSF National AI Research Resource (NAIRR) and related national efforts.

Performance occurs in Salt Lake City, Utah, with a period of performance from July 1, 2026, through June 30, 2028.

Abstract

Safe-AI: sandbox for accelerated feedback and experimentation for AI innovation is a two-year effort to develop and launch a testbed supported by a AI innovation ecosystem built on novel approaches to public-private partnerships. This effort brings together state government, academia, and leading industry partners to drive r & d, foster innovation and adoption, and lay the groundwork for an AI-literate workforce. Further, IT seeks to support both advancements in the broad of adoption of AI in various sectors, with the needs of regulatory compliance in sensitive sectors such as healthcare. The proposal will enable the exploration of novel models for AI adoption in regulated environments across three key use-cases: AI-assisted radiological interpretation, mental health assessment and education, and runtime enforcement infrastructure. This work is led by the safe-AI advisory committee, which includes experts from across the state of utah, including a variety of other academic institutions, government entities, and business leaders (including mindshare ventures, ascend AI labs, and clarion AI partners). The committee will guide the development and deployment of the pilot testbed and its innovation sandboxes, as well as execution of use cases and models for public-private partnerships. The testbed will also engage with related national efforts, including the NSF national AI research resource (nairr). The project will result in the coordination and integration of infrastructure supporting the testbed, formalize and strengthen public-private partnerships, and create strategies for industry to adapt and implement various regulatory schemas in complex sectors in order to reach the market. The testbed safe-AI testbed will leverage and build on AWS innovation sandboxes to provide secure, isolated, and auditable environments for industry leaders and researchers to experiment, demonstrate concepts, and evaluate the results. Together with AWS, the safe-AI innovation sandbox will customize the testbed itself to realize a controlled environment to test AI product readiness to enter and persist in a regulated sector. The activities supported by this award have a significant chance to advance national priorities, bolster regional economic growth, and advance technological progress. Specifically, the development of quantum applications for grid security are highly relevant to U.S. National security, infrastructure resilience, and global competitiveness. Investment in this project is both consistent and well-aligned with the goals and broad strategic objectives of NSF tip. This investment is consistent with agency?s goals and the scientific and technological community?s interest in accelerating the adoption of AI applications in high-impact sectors, including sectors where regulatory frameworks are already pre-existing. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are not planned for this award.