Pacific Northwest National Laboratory and OpenAI partner to accelerate federal permitting

2026-02-26 · Source: OpenAI News · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Digital Government & E-Government, Public Policy & Governance · Depth: Novice, short

Summary

OpenAI has partnered with the U.S. Department of Energy's Pacific Northwest National Laboratory (PNNL) and its PermitAI team to evaluate the potential of coding agents in accelerating federal permitting processes. This collaboration led to the creation of DraftNEPABench, a benchmark designed with 19 subject matter experts to assess AI models on tasks related to National Environmental Policy Act (NEPA) workflows, such as drafting environmental impact statements. Initial findings indicate that generalized coding agents, specifically Codex CLI, could reduce NEPA document drafting time by 1 to 5 hours per subsection, representing up to a 15% reduction in overall drafting time across 18 federal agencies. This suggests a significant advancement in how AI can support complex government operations, particularly in document-heavy regulatory reviews for critical infrastructure projects.

Key takeaway

For federal agencies and project managers navigating complex infrastructure permitting, integrating AI-powered coding agents like those evaluated by PNNL and OpenAI could substantially reduce drafting timelines. You should explore pilot programs for AI assistance in document-heavy regulatory processes to free up human experts for critical judgment and oversight, potentially accelerating project approvals from months to weeks and enhancing U.S. economic competitiveness.

Key insights

AI coding agents can significantly accelerate federal permitting by automating complex document drafting and verification tasks.

Principles

• AI can support complex government workflows.
• Coding agents enhance reasoning model performance.
• Benchmarks are crucial for real-world AI evaluation.

Method

OpenAI and PNNL designed DraftNEPABench, a benchmark for NEPA workflows, using 19 subject matter experts. They evaluated generalized coding agents (Codex CLI) with GPT-5 on tasks requiring document synthesis, fact verification, and structured report drafting via a command-line interface.

In practice

• Use coding agents for regulatory document drafting.
• Implement AI for fact-checking across multiple sources.
• Develop interactive reports from AI-generated content.

Topics

• Federal Permitting
• AI Agents
• National Environmental Policy Act
• Infrastructure Projects
• PermitAI

Best for: Executive, Policy Maker, AI Operations Specialist, Director of AI/ML

Related on AIssential

• White House Pushes AI Tools To Speed Federal Permitting - Broadband Breakfast — AI and automation are accelerating federal environmental reviews and infrastructure permitting.
• Coding Agents Meet Data Science — AI coding agents boost developer productivity but introduce new bottlenecks and require specialized adaptation for data science workflows.
• Can coding agents relicense open source through a “clean room” implementation of code? — AI-assisted code rewrites challenge traditional "clean room" definitions and open-source licensing compliance.
• Dean Ball, on Joining OpenAI: New Power Centers, Frontier AI Policy, & Main Character Energy — Proactive, internal lab engagement is crucial for shaping frontier AI policy and governance.
• Rakuten fixes issues twice as fast with Codex — Integrating AI coding agents like Codex significantly boosts development speed, enhances safety, and fosters autonomous project execution.
• AI agents in research: when productivity comes at the cost of apprenticeship — AI assistants like Claude Code and OpenClaw significantly enhance research efficiency across multiple tasks.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by OpenAI News.