Apr 30, 2024 - Energy & Environment

Department of Energy dives into promises and perils of AI

Illustration of a robot hand adjusting a thermostat.

Illustration: Allie Carl/Axios

The Department of Energy is taking a slew of steps to harness nascent artificial intelligence technologies while preventing an energy crunch, the department announced.

Why it matters: The DOE, with its national labs and supercomputing facilities, is uniquely positioned among government agencies to leap from research to applications.

  • The agency is contemplating putting AI to work to make the electrical grid more productive and efficient, to more effectively anticipate problems, and even to speed up permitting.

Zoom in: The announcements made Monday came in response to an executive order President Biden issued last year.

  • DOE's science advisory board established a working group to focus on the energy needs of data center infrastructure, as well as to convene stakeholders like utilities, to get their take on how to meet the energy demands.
  • According to a DOE statement, the Energy Advisory Board's working group will make recommendations by June on how the U.S. can best meet the growing energy demand for AI, specifically from data centers housing the backbone of generative AI.

Driving the news: In addition, DOE issued a report Monday on how it plans to utilize AI as an agency. It identified near-term opportunities for AI to help in manage grids via planning, permitting, operations and reliability and resilience.

  • "The electrical grid of the United States is among the most complex machines on Earth," the report states.
  • DOE also sees AI as potentially supporting use cases that meet the country's climate goals.
  • It also identifies AI as a tool to improve forecasts of renewable energy production, and for planning networks of charging stations.

Between the lines: For example, on the permitting front, the agency sees large language models as potentially assisting federal permitting compliance and review tasks.

Yes, but: The DOE finds that managing the downsides of AI, including the possibility of damaging the grid or other infrastructure, will be a key priority.

What they found: The DOE report concludes that AI is an opportunity to transform the U.S. grid and contribute to rapid decarbonization, albeit with some risks.

  • "Emerging applications for AI offer the potential to enable change on the grid at a nonlinear pace and scale," the report states.

The intrigue: The department's National Laboratories also issued a companion report on AI and the future of energy on Monday. The study zeroes in on applications for nuclear energy, the grid, carbon management, energy storage and energy materials advancements that could be made within the next decade.

The bottom line: How the U.S. powers itself and makes its grid more reliable and cleaner depends on how the government manages AI.

Go deeper