As artificial intelligence (AI) technologies advance, the applications of AI in Federal agencies are growing as well, expanding beyond pilots to real-world use cases that are heightening operational impact for a variety of Federal missions.
AI is increasing operational efficiency, detecting fraud, and analyzing huge quantities of data at agencies ranging from the U.S. Navy to the Centers for Medicare and Medicaid Services (CMS), AI experts at General Dynamics Information Technology (GDIT) noted in a recent discussion.
Since the Biden administration’s AI executive order last year, Federal agencies have hired more than 150 AI and AI-enabling professionals who are working on critical missions such as guiding efforts to use AI for permitting and advising on government-wide AI investments. This “AI talent surge” reinforces a key point made by the GDIT experts – AI’s impact on the mission is virtually unlimited.
“It almost feels impossible to have a technology discussion without artificial intelligence being mentioned,” said Michael Cole, chief technology officer for GDIT’s Federal civilian business. “There is more to AI for mission support across the Federal government. For instance, AI can support the drive to be more efficient, cost-effective, and more effectively use data.”
Cole’s remarks came as Federal agencies are increasingly integrating AI into daily operations. The U.S. Government Accountability Office recently said 20 agencies reported about 1,200 current or planned AI use cases for missions that include analyzing data from cameras and radar to identify border activities and targeting scientific specimens for planetary rovers.
AI is also expected to help the government defend the nation against increasingly worrisome cyber threats. A GDIT report released last year, in partnership with Splunk, documented the vital role of AI in preemptively identifying and mitigating cyber threats.
Delivering a Conversational AI Solution
Working with the U.S. Navy, GDIT launched a conversational AI solution – known as Amelia – capable of resolving customers’ requests without a live representative. The launch was part of the Navy Enterprise Service Desk (NESD) program, a $136 million knowledge-based service desk solution working to modernize and consolidate more than 90 IT help desks across the Navy into a central service desk.
The Navy “has really understood the power of unleashing artificial intelligence,” said Stephanie Wood, GDIT’s senior program manager on NESD. “Its ability to automate repeatable activities consistently frees up upper tiers and other support staff to do more operationally impacting work. It’s increasing the operational readiness of the Navy and the Department of Defense at large.”
The program represented a huge cultural change for service personnel who were accustomed to speaking with a live agent for support.
But the effort has proven so successful that GDIT is working with other sections of the Navy to integrate Amelia and to implement AIOps on the service branch’s enterprise service desk for automation and analysis.
AI Sifts Masses of Data for Intelligence Agencies
The U.S. intelligence community is also implementing AI and machine learning (ML), using the technologies as a force multiplier to help analysts keep up with masses of incoming data, said Michael Jones, a geospatial solutions expert at GDIT who focuses on intelligence and homeland security.
AI integration enables the intelligence community to apply computer vision and natural language processing to “augment and automate, or assist analysts in their workflows so that a single analyst can now do the workload of an entire team,” Jones added.
One challenge is that AI and ML can require massive compute and storage capabilities, which can result in huge cloud services bills.
To rein in cloud costs, one customer came to GDIT, which provided high-performance computing as a service. “It allowed us to consolidate compute resources and then allow users to onboard and schedule their workloads, even run them over the weekend,” Jones said. “They utilized compute when it wasn’t being used for mission-critical use cases.”
AI Finds Healthcare Fraud
At CMS, the AI focus is on detecting fraud. GDIT has worked with CMS officials to implement complex fraud detection systems utilizing AI and ML that are demonstrating measurable results, said Andrew Needs, senior manager for fraud analytics at GDIT, who focuses on the Federal healthcare space.
Taking time to gain trust in rapidly advancing AI technologies is critical for success. GDIT took a transparent approach to gain that trust, gradually replacing rule-based fraud detection algorithms with more effective AI and machine learning algorithms.
“Transparency has been very important,” Needs said. “When people think of AI, they might think of a black box around a lot of these detection algorithms. We really want to remove that for our customers and make them feel comfortable with these really advanced methods.”
Going forward, the potential uses of AI for Federal mission success are “really only up to our imagination. There’s no ceiling on where we’re going to go,” Cole added.
For more insights on AI for Federal missions, view the entire discussion.