The U.S. Department of Energy’s Office of Cybersecurity, Energy Security, and Emergency Response, in collaboration with Lawrence Livermore National Laboratory, has launched the Mjölnir AI Testbed, a platform designed to help utilities and technology providers evaluate the security and reliability of AI models used in energy systems.
The initiative supports the broader Genesis Mission, under which DOE recently announced $293 million in funding for 26 science and technology challenges, including securing AI-driven tools.
While AI can enhance energy operations, officials warn that the technology also introduces new vulnerabilities, including risks of manipulation and data exposure.
The Mjölnir Testbed allows users to upload AI models and run adversarial tests to assess resilience against cyberattacks. It evaluates whether models may behave incorrectly, expose sensitive data, or fail under compromised conditions, offering insights into their overall robustness.
“The new platform is designed to support a broad range of energy sector stakeholders, including electric utilities, grid operators, energy technology vendors, and AI solution providers, national laboratories, research institutions, and other qualified organizations,” DOE said in an official update.
By enabling side-by-side comparisons, the platform helps stakeholders identify safer AI solutions and make informed deployment decisions.