
Some Federal agencies are already taking steps to comply with the Trump administration’s Artificial Intelligence Action Plan and other executive orders, including deploying AI systems – such as those that employ agentic AI – with the goal of creating more tailored approaches to the technology.
The Trump administration’s AI Action Plan released late last month laid the groundwork and set policy objectives for Federal agencies to accelerate their development and deployment of AI systems.
Some agencies – including the Department of Energy (DoE) and the National Oceanic & Atmospheric Administration (NOAA) – have already begun breaking down the action plan and other AI-related executive orders to carry out their priorities.
One system recently developed by DoE alerts department officials to how executive orders impact them.
That system, coined the “EO Encompass” system, uses chatbots and grounding technology to gather user feedback and deliver timely updates on executive orders, said Bridget Carper, the agency’s deputy chief information officer for architecture, engineering, technology and innovation, while speaking at a Leadership Connect webinar on Aug. 14.
The system was created after non-policy staff struggled to track rapidly changing directives, and automatically alerts relevant teams to new orders, their impacts, and provides deadlines.
“It gives you that update, so then you’re able to adapt to see ‘wait we have something due in like three months,’ or ‘we have something due in 30 days,’” explained Carper. “It gives you, like that flag.”
The AI Action Plan and related executive orders in particular have been driving internal conversations on moving forward with AI initiatives, Carper said, adding that a key point of discussion has been “removing barriers,” especially regarding permitting for AI data centers.
“We have developed a tool around decreasing the timeline of … permitting, but also ensuring that the environmental impacts are noted,” said Carper, adding that in the future there are more “policies coming out as well.”
Those data centers are a priority under the AI Action Plan and other executive orders which have called for AI data centers to be built in most states.
Pushes for government to employ more AI have also led to the adoption of agentic AI systems, which Monica Youngman, chief scientist at the National Weather Service at NOAA, said she has embraced with the mindset of if “there’s low risk, let’s try it, we have more to gain from moving forward into that space than there is to lose.”
One system within NOAA uses agentic AI but is overseen by a human making the final decisions. The system uses “a system of agentic AI that decides what data to look at to verify a report and makes those decisions on how to provide that back to the user and make that evaluation.”
“We are starting to use agentic AI in basically controlled systems where there’s still a human envelope at the end,” said Youngman.
NOAA also has made Gemini AI services available throughout its entire workforce, Youngman added, saying that they worked with an existing contract to allow all new users to access Gemini within the agency’s workspace.
“We started with a group of about 200 testers across NOAA, got feedback – we were able to document what are those use cases that work very well versus don’t work very well,” explained Youngman, adding that the agency also documented privacy and data concerns as part of that pilot. “It’s been such a game changer, with lots of innovation, new ideas.”
Youngman said she is also “very excited for the potential for ChatGPT” following the General Services Administration’s OneGov contract announcement with the AI service provider to allow agencies to access its platform.