White House Deploys AI to Speed Environmental Permitting

The Trump administration is spending $14 million at three federal agencies for AI tools that can help speed up the environmental permitting process.

The outlay can be seen as an extension of both the White House’s desire to get more roads, bridges, mines, and data centers built faster, and its broad campaign to expand the federal government’s use of AI to increase efficiency.

The new tools by design don’t take advantage of AI’s full cognitive or reasoning capabilities, but the Federal Permitting Improvement Steering Council still thinks they can dramatically cut wait times and snip out human errors, Emily Domenech, the agency’s executive director, said in a June 18 interview.

“Some of the stuff we’re talking about automating is not that complicated,” Domenech said. “It’s stuff we’ve had the ability to do for quite some time, like digital forms for applications, as opposed to a PDF. We’ve been able to do digital forms using just a basic Google function for a decade. And the federal government is just really behind.”

Under a June 12 announcement, the US Army Corps of Engineers, US Coast Guard, and National Telecommunications and Information Administration will get funding from FPISC to make their permitting processes more efficient.

Agency Uses

The Army Corps’ $7.1 million will use AI to develop a workflow system that can process environmental reviews in a unified way, knitting together procedural steps that are currently fragmented, and automating many routine tasks that now require human time and effort, Domenech said.

The Corps believes the system can save more than 42,000 hours a year, or the equivalent of 21 full-time employees, she said.

The Coast Guard will use its $4.8 million to create a new online bridge permit application system to replace its current one, which the agency thinks requires too much copying and pasting from applicants’ forms, Domenech said.

The agency will use AI to automatically scan and upload documents into its administrative record system and incorporate its internal geographic data into a project’s file from the start.

“These are all things that had to be done manually for every single bridge permit, up until this point,” Domenech said. “It’s also, frankly, producing errors in the permits, because they’re inevitably going to have those, as you have a manual copy and paste.”

The new system will also be able to scan existing environmental and navigational documents to understand the impacts of mitigation options on other, related projects, Domenech said. That could save applicants and their consultants hundreds of hours of work in comparing the options from scratch, she said.

At NTIA, $2.8 million in funding will build on a previous AI investment from the Biden administration to make smarter decisions about whether a project requires an environmental assessment (EA) or the more detailed environmental impact statement.

The system will also aim to use AI to automate drafting pieces of the EA that are “always the same, so that we’re not drafting them by hand every single time,” Domenech said.

NTIA—which, among other things, helps improve broadband access across the US—has said the AI tool could reduce its EA timelines by three to 60 days and cut costs by up to 50%, according to Domenech.

Future Efforts

Going forward, FPISC expects even more agencies to adopt AI tools to improve their permitting decisions.

“We’re really looking for proposals that could show us real time and cost savings, real opportunities to move from manual systems to AI-driven systems, and ways to really maximize the FTEs we have to do the hard, complicated reviews that can’t be done through an AI system, or to be checking the work of the AI system instead of doing all the work from scratch,” Domenech said.

She also said she doesn’t fear the threat of increased litigation on the grounds that a permitting decision was made by AI, because the Trump administration isn’t using AI to replace permit authorizers.

“They’re not replacing the FTE review,” Domenech said. “They’re not even replacing much of the work that goes into the technical work that’s done by that subject matter expert. They’re replacing the mundane tasks, like somebody literally copying and pasting something from one spreadsheet to another.”