Despite this grim prognosis, Congress could still improve implementation of a key permitting barrier, the National Environmental Policy Act, by fixing the federal government’s broken systems for managing and sharing NEPA documentation and data. These opaque and incompatible systems frustrate essential interagency coordination, contributing immeasurably to NEPA’s delays and frustrations. But it’s a problem with clear, available, workable solutions — and at low political cost.
Both of us saw these problems firsthand. Marc helped manage NEPA implementation at the Environmental Protection Agency, observing the federal government’s slow and often flailing attempts to use technology to improve internal agency processes. Elizabeth, meanwhile, spent two years overcoming NEPA’s atomized data ecosystem to create a comprehensive picture of NEPA litigation.
Even so, it’s difficult to illustrate the scope of the problem without experiencing it. Some agencies have bespoke systems to house crucial and unique geographic information on project areas. Other agencies lack ready access to that information, even as they examine project impacts another agency may have already studied. Similarly, there is no central database of scientific studies undertaken in support of environmental reviews. Some agencies maintain repositories for their environmental assessments — arduous but less intense environmental reviews than the environmental impact statements NEPA requires when a federal agency action substantially impacts the environment. But there’s still no unified, cross-agency EA database. This leaves agencies unable to efficiently find and leverage work that could inform their own reviews. Indeed, agencies may be duplicating or re-duplicating tedious, time-consuming efforts.
NEPA implementation also relies on interagency cooperation. There, too, agencies’ divergent ways of classifying and communicating about project data throws up impediments. Agencies rely on arcane data formats and often incompatible platforms. (For the tech-savvy, an agency might have a PDF-only repository while another has XML-based data formats.) With few exceptions, it’s difficult for cooperating agencies to even know the status of a given review. And it produces a comedy of errors for agencies trying to recruit and develop younger, tech-savvy staff. Your workplace might use something like Asana or Trello to guide your workflow, a common language all teams use to communicate. The federal government has a bureaucratic Tower of Babel.
Yet another problem, symptomatic of inadequate transparency, is that we have only limited data on the thousands of NEPA court cases. To close the gap, we sought to understand — using data — just how sprawling and unwieldy post-review NEPA litigation had become. We read every available district and appellate opinion that mentioned NEPA from 2013 to 2022 (over 2,000 cases), screened out those without substantive NEPA claims, and catalogued their key characteristics — plaintiffs, court timelines and outcomes, agencies, project types, and so on. Before we did this work, no national NEPA litigation database provided policymakers with actionable, data-driven insights into court outcomes for America’s most-litigated environmental statute. But even our painstaking efforts couldn’t unearth a full dataset that included, for example, decisions taken by administrative judges within agencies.
We can’t manage what we can’t measure. And every study in this space, including ours, struggles with this type of sample bias. Litigated opinions are neither random nor representative; they skew toward high-stakes disputes with uncertain outcomes and underrepresent cases that settle on clear agency error or are dismissed early for weak claims. Our database illuminates litigation patterns and timelines. But like the rest of the literature, it cannot offer firm conclusions about NEPA’s effectiveness. We need a more reliable universe of all NEPA reviews to have any chance — even a flawed one — at assessing the law’s outcomes.
In the meantime, NEPA policy debates often revolve unproductively around assumptions and anecdotes. For example, Democrats can point to instances when early and robust public engagement appeared essential for bringing projects to completion. But in the absence of hard data to support this view, GOP reformers often prefer to limit public participation in the name of speeding the review process. The rebuttal to that approach is persuasive: Failing to engage potential project opponents on their legitimate concerns merely drives them to interfere with the project outside the NEPA process. Yet this rebuttal relies on assumptions, not evidence. Only transparent data can resolve the dispute.
Some of the necessary repair work is already underway at the Council on Environmental Quality, the White House entity that coordinates and guides agencies’ NEPA implementation. In May, CEQ published a “NEPA and Permitting Data and Technology Standard” so that agencies could voluntarily align on how to communicate NEPA information with each other. Then in June, after years using a lumbering Excel file containing agencies’ categorical exclusions — the types of projects that don’t need NEPA review, as determined by law or regulation — CEQ unveiled a searchable database called the Categorical Exclusion Explorer. The Pacific Northwest National Laboratory’s PermitAI has leveraged the EPA’s repository of environmental impact statements and, more recently, environmental review documents from other agencies to create an AI-powered queryable database. The FAST-41 Dashboard has brought transparency and accountability to a limited number of EISs.
But across all these efforts, huge gaps in data, resources, and enforcement authority remain. President Trump has issued directives to agencies to speed environmental reviews, evincing an interest in filling the gaps. But those directives don’t and can’t compel the full scope of necessary technological changes.
Some members of Congress are tuned in and trying to do something about this. Representatives Scott Peters, a Democrat from California, and Dusty Johnson, Republican of South Dakota, deserve credit for introducing the bipartisan ePermit Act to address all of these challenges. They’ve identified key levers to improve interagency communication, track litigation, and create a common and publicly accessible storehouse of NEPA data. Crucially, they recognize the make-or-break role of agency Chief Information Officers who are accountable for information security. Our own attempts to upgrade agency technology taught us that the best way to do so is by working with — not around — CIOs who have a statutory mandate.
The ePermit Act would also lay the groundwork for more extensive and innovative deployment of artificial intelligence in NEPA processes. Despite AI’s continuing challenges around information accuracy and traceability, large language models may eventually be able to draft the majority of an EIS on their own, with humans involved to oversee.
AI can also address hidden pain points in the NEPA process. It can hasten the laborious summarization and incorporation of public comment, reducing the legal and practical risk that agencies miss crucial public feedback. It can also help determine whether sponsor applications are complete, frequently a point of friction between sponsors and agencies. AI can also assess whether projects could be adapted to a categorical exclusion, entirely removing unnecessary reviews. And finally, AI tools are a concession to the rapid turnover of NEPA personnel and depleted institutional knowledge — an acute problem of late.
Comprehensive, multi-agency legislation like the ePermit Act will take time to implement — Congress may want or even need to reform NEPA before we get the full benefit of technology improvements. But that does not diminish the urgency or value of this effort. Even Representative Jared Huffman of California, a key Democrat on the House Natural Resources Committee with impeccable environmental credentials, offered words of support for the ePermit Act, while opposing other NEPA reforms.
Regardless of what NEPA looks like in coming years, this work must begin at some point. Under every flavor of NEPA reform, agencies will need to share data, coordinate across platforms, and process information. That remains true even as court-driven legal reforms and Trump administration regulatory changes wreak havoc with NEPA’s substance and implementation. Indeed, whether or not courts, Congress, or the administration reduce NEPA’s reach, even truncated reviews would still be handicapped by broken systems. Fixing the technology infrastructure now is a way to future-proof NEPA.
The solution won’t be as simple as getting agencies to use Microsoft products. It’s long past time to give agencies the tools they need — an interoperable, government-wide platform for NEPA data and project management, supported by large language models. This is no simple task. To reap the full benefits of these solutions will require an act of Congress that both provides funding for multi-agency software and requires all agencies to act in concert. This mandate is necessary to induce movement from actors within agencies who are slow to respond to non-binding CEQ directives that take time away from statutorily required work, or those who resist discretionary changes to agency software as cybersecurity risks, no matter how benign those changes may be. Without appropriated money or congressional edict, the government’s efforts in this area will lack the resources and enforcement levers to ensure reforms take hold.
Technology improvements won’t cure everything that ails NEPA. This bill won’t fix the deep uncertainty unleashed by the legal chaos of the last year. But addressing these issues is a no-regrets move with bipartisan and potentially even White House support. Let it be done.
Guyanese President Irfaan Ali. Riccardo Savi/Getty Images for Concordia Summit