Building a Comprehensive NEPA Database to Facilitate Innovation
The Inflation Reduction Act and the Infrastructure Innovation and Jobs Act are set to drive $300 billion in energy infrastructure investment by 2030. Without permitting reform, lengthy review processes threaten to make these federal investments one-third less effective at reducing greenhouse gas emissions. That’s why Congress has been grappling with reforming the National Environmental Policy Act (NEPA) for almost two years. Yet, despite the urgency to reform the law, there is a striking lack of available data on how NEPA actually works. Under these conditions, evidence-based policy making is simply impossible. With access to the right data and with thoughtful teaming, the next administration has a golden opportunity to create a roadmap for permitting software that maximizes the impact of federal investments.
Challenge and Opportunity
NEPA is a cornerstone of U.S. environmental law, requiring nearly all federally funded projects—like bridges, wildfire risk-reduction treatments, and wind farms—to undergo an environmental review. Despite its widespread impact, NEPA’s costs and benefits remain poorly understood. Although academics and the Council on Environmental Quality (CEQ) have conducted piecemeal studies using limited data, even the most basic data points, like the average duration of a NEPA analysis, remain elusive. Even the Government Accountability Office (GAO), when tasked with evaluating NEPA’s effectiveness in 2014, was unable to determine how many NEPA reviews are conducted annually, resulting in a report aptly titled “National Environmental Policy Act: Little Information Exists on NEPA Analyses.”
The lack of comprehensive data is not due to a lack of effort or awareness. In 2021, researchers at the University of Arizona launched NEPAccess, an AI-driven program aimed at aggregating publicly available NEPA data. While successful at scraping what data was accessible, the program could not create a comprehensive database because many NEPA documents are either not publicly available or too hard to access, namely Environmental Assessments (EAs) and Categorical Exclusions (CEs). The Pacific Northwest National Laboratory (PNNL) also built a language model to analyze NEPA documents but contained their analysis to the least common but most complex category of environmental reviews, Environmental Impact Statements (EISs).
Fortunately, much of the data needed to populate a more comprehensive NEPA database does exist. Unfortunately, it’s stored in a complex network of incompatible software systems, limiting both public access and interagency collaboration. Each agency responsible for conducting NEPA reviews operates its own unique NEPA software. Even the most advanced NEPA software, SOPA used by the Forest Service and ePlanning used by the Bureau of Land Management, do not automatically publish performance data.
Analyzing NEPA outcomes isn’t just an academic exercise; it’s an essential foundation for reform. Efforts to improve NEPA software have garnered bipartisan support from Congress. CEQ recently published a roadmap outlining important next steps to this end. In the report, CEQ explains that organized data would not only help guide development of better software but also foster broad efficiency in the NEPA process. In fact, CEQ even outlines the project components that would be most helpful to track (including unique ID numbers, level of review, document type, and project type).
Put simply, meshing this complex web of existing softwares into a tracking database would be nearly impossible (not to mention expensive). Luckily, advances in large language models, like the ones used by NEPAccess and PNNL, offer a simpler and more effective solution. With properly formatted files of all NEPA documents in one place, a small team of software engineers could harness PolicyAI’s existing program to build a comprehensive analysis dashboard.
Plan of Action
The greatest obstacles to building an AI-powered tracking dashboard are accessing the NEPA documents themselves and organizing their contents to enable meaningful analysis. Although the administration could address the availability of these documents by compelling agencies to release them, inconsistencies in how they’re written and stored would still pose a challenge. That means building a tracking board will require open, ongoing collaboration between technologists and agencies.
- Assemble a strike team: The administration should form a cross-disciplinary team to facilitate collaboration. This team should include CEQ; the Permitting Council; the agencies responsible for conducting the greatest number of NEPA reviews, including the Forest Service, Bureau of Land Management, and the U.S. Army Corps of Engineers; technologists from 18F; and those behind the PolicyAI tool developed by PNNL. It should also decide where the software development team will be housed, likely either at CEQ or the Permitting Council.
- Establish submission guidelines: When handling exorbitant amounts of data, uniform formatting ensures quick analysis. The strike team should assess how each agency receives and processes NEPA documents and create standardized submission guidelines, including file format and where they should be sent.
- Mandate data submission: The administration should require all agencies to submit relevant NEPA data annually, adhering to the submission guidelines set by the strike team. This process should be streamlined to minimize the burden on agencies while maximizing the quality and completeness of the data; if possible, the software development team should pull data directly from the agency. Future modernization efforts should include building APIs to automate this process.
- Build the system: Using PolicyAI’s existing framework, the development team should create a language model to feed a publicly available, searchable database and dashboard that tracks vital metadata, including:
- The project components suggested in CEQ’s E-NEPA report, including unique ID numbers, level of review, document type, and project type
- Days spent producing an environmental review (if available; this information may need to be pulled from agency case management materials instead)
- Page count of each environmental review
- Lead and supporting agencies
- Project location (latitude/longitude and acres impacted, or GIS information if possible)
- Other laws enmeshed in the review, including the Endangered Species Act, the National Historic Preservation Act, and the National Forest Management Act
- When clearly stated in a NEPA document, cost of producing the review
- When clearly stated in a NEPA document, staff hours used to produce the review
- When clearly stated in a NEPA document, jobs and revenue created by the project
- When clearly stated in a NEPA document, carbon emissions mitigated by the project
Conclusion
The stakes are high. With billions of dollars in federal climate and infrastructure investments on the line, a sluggish and opaque permitting process threatens to undermine national efforts to cut emissions. By embracing cutting-edge technology and prioritizing transparency, the next administration can not only reshape our understanding of the NEPA process but bolster its efficiency too.
This action-ready policy memo is part of Day One 2025 — our effort to bring forward bold policy ideas, grounded in science and evidence, that can tackle the country’s biggest challenges and bring us closer to the prosperous, equitable and safe future that we all hope for whoever takes office in 2025 and beyond.
It’s estimated that only 1% of NEPA analyses are Environmental Impact Statements (EISs), 5% are Environmental Assessments (EAs), and 94% are Categorical Exclusions (CEs). While EISs cover the most complex and contentious projects, using only analysis of EISs to understand the NEPA process paints an extremely narrow picture of the current system. In fact, focusing solely on EISs provides an incomplete and potentially misleading understanding of the true scope and effectiveness of NEPA reviews.
The vast majority of projects undergo either an EA or are afforded a CE, making these categories far more representative of the typical environmental review process under NEPA. EAs and CEs often address smaller projects, like routine infrastructure improvements, which are critical to the nation’s broader environmental and economic goals. Ignoring these reviews means disregarding a significant portion of federal environmental decision-making; as a result, policymakers, agency staff, and the public are left with an incomplete view of NEPA’s efficiency and impact.