EPA Outlines New AI Strategy and Governance Framework
EPA plans to accelerate the development and deployment of artificial intelligence (AI) throughout the agency’s operations, according to two documents released in October 2025.
The documents—an AI strategy and associated “compliance plan”—come in response to the Trump administration’s Executive Order 14179: Removing Barriers to American Leadership in Artificial Intelligence and a related memorandum from the Office of Management and Budget (OMB).
“EPA is investing in its overall architecture to deploy more robust AI tools. These investments include data fabric, cloud infrastructure, network infrastructure, and additional security, operational, financial, and post deployment monitoring tools,” the compliance plan states.
According to the AI strategy, EPA is investigating “Expanding a pilot into a coding-focused generative AI tool” for software development, “Procuring more American Generative AI tools,” and “Expanding Generative AI capabilities in its cloud vendors.”
Current AI projects include a machine learning model that predicts pollution exposure pathways, a model that assists with records scheduling, and a model to aid in facility inspections, according to the AI strategy. In addition, in May 2025, EPA introduced an internal generative AI chat tool for employee use.
Neither document discusses whether AI may be used to speed new chemical or pesticide reviews, though EPA Administrator Lee Zeldin noted that it was a possibility earlier this year.
AI Governance and Risk Management
To comply with OMB’s memo, EPA established a political-level AI Governance Board chaired by the Deputy Administrator, which will oversee and coordinate the responsible use of AI across the agency. The board is assisted by a subcommittee with “delegates from the many diverse offices across EPA,” the AI strategy states.
EPA says these governance structures will implement risk-mitigation controls for “high-impact” AI use cases. Additional risk-management efforts include employee training, development of a “Generative AI Rules of Behavior” guide, and ongoing monitoring and feedback cycles for AI systems.
The agency also emphasized its commitment to transparency, pointing to its practice of making code publicly available on platforms such as Data.gov and GitHub. According to the compliance plan, “EPA’s GitHub site is one of the more popular federal government GitHub repositories for staff development, collaboration, and code sharing.”
More on EPA’s active projects, and other AI use cases, can be found on EPA’s AI Use Case Inventory.
