Federal AI Use Cases Highlight Adoption Opportunities, Challenges

Federal agencies are leveraging artificial intelligence to uncover opportunities and risks as leaders continue to integrate the technology to advance their missions.

Top tech officials gathered at the inaugural GovCIO Media & Research AI Summit in Tysons Corner, Virginia, Thursday to discuss the security and data challenges surrounding AI and how they are planning to integrate this technology in the years to come. Here is a look at some of those highlighted use cases.

TMF Investments Spur AI Projects

The General Services Administration (GSA)’s Technology Modernization Fund (TMF) is helping shape initial AI projects across government since President Biden’s executive order on the safe, secure and trustworthy development and use of AI. GSA’s call for proposals on AI returned 40 teams representing agencies at varying stages of readiness and needs, including prototyping AI systems, integrating workflows and more, said TMF Acting Executive Director Jessie Posilkin.

She cited the Department of Homeland Security (DHS)’s award of $27 million in TMF funds for its Homeland Security and Information Network modernization to enable AI capabilities that enhance public safety missions.

“One pilot is focused on classifying and tagging metadata for improved information organization, and the other is developing large language models to generate content summaries, which are both really needed for processing the huge amount of information that they have,” said Posilkin. “They are exemplifying some of the best practices in AI adoption that we want to see, using iterative testing, using controlled sandboxes for experimentation, having strong governance and, even more importantly, validating their use cases through consistent user engagement, ensuring that their tools serve specific, well-defined operational needs.”

GSA is also helping the Social Security Administration (SSA) reduce the wait time in their disability claims processing system. Posilkin said many Americans with disabilities have died before getting a decision about their benefits or accessing the benefits they need.

“That kind of wait time has huge implications for what real people experience,” said Posilkin.

Air Force Eyes AI to Support Acquisition

Alexis Bonnell, CIO and director of the Digital Capabilities Directorate with the Air Force Research Laboratory (ARFL), noted AI’s role in the human journey. AFRL discovered that employees need an ally when it comes to technology and data adoption in an AI tool supporting its acquisition professionals.

“What [AI] did had to be helpful to people, and it had to be done out of the respect for that person. If you are one of our amazing acquisitions people, we’ve really [tried] to make sure that you can do the important things right,” said Bonnell.

The lab employed a retrieval augmented generation (RAG) approach to enhance the accuracy and reliability of generative AI models, which is enabling the Air Force to meet its mission.

“Instead of having and querying all the collective models associated knowledge, it means that that person gets to say, ‘I know what is relevant to me and I know the information that matters,'” said Bonnell. “That might be in emails, it might be in PDFs, it might be in a database, but actually, what we see them do is trying to create an intimate universe and a knowledge universe for themselves.”

AI to Support the Economy

According to Eric Smith, acting deputy assistant secretary for policy at the Economic Development Administration (EDA), AI is playing a role in its Tech Hubs program enacted as part of the CHIPS and Science Act of 2022.

The initiative is focused on bringing economic growth and global economic competitiveness throughout the country, especially in photonic sensing, quantum-enabling technologies and biotechnology. Smith highlighted how AI is critical to enabling nearly every one of the program’s technology areas.

“I think our hubs are doing a great job of identifying how to do that, how to use AI to accelerate innovation and innovation-based production and identifying some of the gaps where maybe we’re not able to use it effectively yet, but we can build up the capability of doing so,” said Smith.

Solving the AI Challenge

According to Jim Gwiazda, vice president of the public sector at Equinix, one challenge with AI is regarding the physical location of the data. AI can’t exist without a place to live, whether that be on-prem, in the cloud or within a data center and at scale.

“AI is not confined to a single process in a single location,” said Gwiazda. “It encompasses a series of interconnected workloads distributed across various platforms and various locations, and determining which workload should be placed where requires careful consideration as part of that.”

The officials noted the promise of AI as agencies work around the ethics for its responsible use and scaling it.

“This is the first technology that allows us to put all our knowledge on the table,” Bonnell added. “It allows us to have an intimate relationship with it.”

“There’s enormous potential here to transform American government services further. Half of our recent proposals that we just got in our most recent cohort were AI-related,” added Posilkin. “Clearly this is not something that is going away. It isn’t a flash in the pan, and we’re seeing that AI is just one piece of the way that civilian agencies and defense agencies need to be thinking about improving their missions.”