DOE is Advancing the AI Innovation Ecosystem

Artificial intelligence (AI) is transforming science and technology, with profound implications for America and the U.S. Department of Energy's (DOE) science, energy, and security mission. DOE already leverages AI for activities such as advanced computing, emergency response, environmental modeling and climate forecasting, and materials research. AI holds the potential to enable streamlined environmental permitting, deployment of fusion power at scale, and improved electrical grid reliability. DOE is focused on developing the tools necessary to advance the science of AI and develop a world-leading AI innovation ecosystem.

For decades, DOE and scientists, engineers, and technologists at its national laboratories have been advancing the development of AI. The 2023 Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence calls on DOE to lead (i) the development of tools for building AI foundation models for science, (ii) catalyzing development of AI tools to mitigate climate changes risks, and (iii) the expansion of partnerships to build AI foundation models for science, energy, and security.

Foundation models are developed and trained on broad data inputs and can then be adapted to a wide range of different tasks. Foundation models have the ability to develop insight and discover meaningful patterns in vast troves of data that may initially seem unmanageable. Foundation models built specifically for DOE’s science, energy, and security mission hold promise for transforming both the way the Department executes its mission and the impact and reach of new science discoveries and applications on everything from critical minerals to climate forecasting to nuclear nonproliferation.

The DOE complex has produced a number of AI foundation models. In one example, DOE’s National Renewable Energy Laboratory demonstrated that an AI foundation model can use DOE national laboratories’ vast technical data and reports to answer users’ questions about Puerto Rico’s transition to 100% renewable energy. In another example, Los Alamos National Laboratory is developing AI foundation models for biological research in order to study viral protein sequences using large language models, predict properties of mutated and new protein sequences, generate or repurpose existing drugs, and investigate how to provide new guardrails for biosecurity research.

DOE, the DOE national laboratories, and our partners are powering a robust AI ecosystem by advancing secure, trustworthy, and equitable AI for all Americans. This web portal highlights a non-exhaustive selection of DOE AI tools, foundation models, and partnerships for applications in science, energy, climate, and security.