AI and the federal permit process: Understanding the next steps

Artificial Intelligence (AI) is being touted as either a tool that can be used to

help make things easier for humanity, or as the future Skynet destroying civilization. 

Applying AI to help the US Army Corps of Engineers (Corps) improve wetland permit application processing will likely not result in the destruction of civilization, but an incorrect permit, maybe processed using AI, could lead to further erosion of trust in the federal government.  

Perhaps more alarming, most voices clamoring for increased use of AI tools do not mention cybersecurity at all.  Imagine Skynet being controlled by hackers. Now imagine hackers issuing permits to fill wetlands. 

Most of us have seen or heard the stories about “AI hallucinations,” or nonexistent facts AI somehow invents. These errors emerge when an AI tool does not have enough data or is not trained properly and makes up responses. 

I am proposing two strategic options for the implementation of AI for permit application processing that can avoid these pitfalls and improve accuracy and results.

The first option is to apply AI in an iterative manner to wetland permitting, using existing regulations and processing procedures. The benefit of using the Corps as a test case is the massive database of permits and correspondence that could be used to train the AI module by a team of experts.  

There are three areas that could benefit from this approach: 

• Assistance with Permit Applications. Provide guided walkthrough for new applicants to navigate the maze of Nationwide and Regional General Permits, assisting with options to meet conditions and reduce impacts. It would also help with more complex permits, clarifying what is needed for a Public notice and complete application. It would also populate the database and provide documentation of “non-reporting Nationwide Permits” to applicants.

• Assistance with NEPA documentation. AI could augment the existing Word macro (automatically loads information) to develop comprehensive NEPA documents with minimal editing.

• Provide a first response to FOIA requests. This is a time-consuming duty that could be made less onerous and improve response time to the public.

Although the above approaches use existing regulations and can be quickly impactful, many of the regulations were written in the late 1970’s. To ensure lasting and impactful change, a more holistic approach needs to take place. 

To achieve transformative change in the permit process, a dedicated team would have to start with the Clean Water Act itself and design a new permitting process. The existing regulations written in the 1970’s would be rewritten to align the requirements in the CWA law itself integrating modern AI techniques using existing information in the Corps’ permit database to drive transformative change in the process with goals of improving transparency in the process and decreasing permit processing time.  

A complete overhaul of the process to meet the goals and objectives of the Clean Water Act using AI to streamline approvals would be truly transformative and support the Nation’s goals to balance wetland protection with reasonable development.  

Mark Sudol

Senior Advisor

Prior to joining Dawson & Associates, Mark was Chief of the U.S. Army Corps of Engineers Regulatory Program and oversaw the Corps’ regulatory responses to U.S. Supreme Court’s Rapanos and Northern Cook County decisions. He was also senior lead for the team that wrote the Mitigation Banking Regulations.