Twenty-seven FBI field offices were working fraud cases that looked local. Then Washington finally put the data in the same room.
What emerged wasn't dozens of unrelated problems. The FBI says investigators connected over 40 cases to the same strategic threat and the same group of overseas actors.
From the FBI:
Launched in August, the NFDC brings together the FBI, federal law enforcement agencies, Offices of Inspector General, and state partners in a first-of-its-kind effort to identify fraud across taxpayer-funded programs. Rather than evaluating investigations in isolation, the center combines cross-agency data and artificial intelligence to help investigators spot patterns, connect cases, and dismantle organized criminal networks.
The initiative stems from a broader "whole of government" approach mandated by Executive Order 14395, which established the White House Task Force to Eliminate Fraud.
The discovery came through the new National Fraud Detection Center, launched in August. The center brings together information from the FBI, Homeland Security Investigations, IRS Criminal Investigation, Treasury, FinCEN, inspectors general, state agencies, and others.
Instead of examining one suspicious student loan, Medicare payment, or SBA loan at a time, investigators can search for connections across programs and jurisdictions. The center uses data analytics and AI to help spot patterns that individual offices might miss.
Related: Is Karen Bass Headed for Reelection or Prison?
The student-loan investigation offers the clearest example. Individual cases appeared small enough that some might never have attracted major federal attention. Once the data was combined, investigators concluded they were looking at pieces of a much larger foreign-directed operation.
Fraudsters had figured out something the federal government has struggled with for years: agencies don't always talk to each other.
The Government Accountability Office reported in June that agencies can use more than 100 federal data sources to verify whether people qualify for benefits or payments. GAO examined nine major sources and found data-quality problems in every one. It also found no government-wide rules ensuring those systems can efficiently share information.
The price of those weaknesses isn't theoretical.
Federal agencies reported an estimated $186 billion in improper payments during fiscal 2025. Improper payments include errors and overpayments, so they shouldn't automatically be called fraud. GAO separately estimates that actual fraud costs the federal government somewhere between $233 billion and $521 billion annually, based on data from 2018 through 2022.
Those numbers help explain why connecting databases matters.
A $200,000 scheme in one jurisdiction may look like a modest federal case. Similar cases spread across 20 states can reveal something entirely different. The FBI says the new system allows investigators to see when smaller losses are attached to organized networks operating across state lines or from overseas.
AI can help find those connections, but it can't rescue bad information.
GAO has repeatedly warned that artificial intelligence depends on reliable data. Inconsistent, incomplete, or poorly managed records can produce bad leads just as easily as good ones. Better technology doesn't erase the need for accurate government records and careful investigators.
For years, Americans have been told federal programs are simply too large and complicated to police perfectly.
Maybe the better question is how much fraud survived because Washington built enormous databases and never made them work together.
Forty investigations across 27 FBI offices suggest we are beginning to find out.