Pacific Northwest National Laboratory and OpenAI partner to accelerate federal permitting
OpenAI Blog
February 26, 2026
AI-Generated Deep Dive Summary
OpenAI and the Pacific Northwest National Laboratory (PNNL) have launched a new benchmark called DraftNEPABench to evaluate how AI coding agents can accelerate federal permitting processes. This initiative focuses on improving efficiency in drafting documents related to the National Environmental Policy Act (NEPA), which is crucial for evaluating the environmental impact of major infrastructure projects. By leveraging AI models, the partnership aims to reduce NEPA document drafting time by up to 15%, a significant step toward modernizing the often slow and costly permitting process.
Federal permitting, particularly under NEPA, involves extensive reviews of technical reports, regulatory compliance, and structured documentation. The collaboration with 19 subject matter experts highlights the potential of AI agents like Codex CLI to handle complex tasks such as reading hundreds of pages of technical content, verifying facts across multiple sources, and drafting detailed analyses. These tools are designed to assist human reviewers by speeding up the process while maintaining accuracy and meeting legal standards.
The benchmark evaluates AI models on their ability to perform tasks spanning NEPA document sections from various federal agencies. Early findings suggest that generalized coding agents can save 1 to 5 hours per subsection, depending on the complexity of the task. This reduction in time not only accelerates project approvals but also lowers costs and brings benefits to communities faster.
For readers interested in AI's role in government processes, this collaboration underscores the growing potential of AI to transform traditional workflows. By integrating AI into NEPA reviews, agencies can streamline operations while ensuring compliance with regulations. The use of dynamic, web-based reports and interactive visualizations further enhances human validation processes, paving the way for more efficient and transparent decision-making.
This initiative is part of a broader push to modernize how the U.S. government handles critical infrastructure projects. As AI systems become more sophisticated, their ability to assist in areas like civil engineering, environmental analysis, and regulatory compliance will play a key role in building a faster, safer, and more competitive economy. The success of DraftNEPABench represents a meaningful step forward in leveraging AI responsibly to support complex government workflows.
Verticals
airesearch
Originally published on OpenAI Blog on 2/26/2026