Rising fraud and shrinking staff signal an urgent need for AI in government
Have you ever opened your wallet and felt panicked? You count the bills once, twice and still come up short. You rummage through every pocket, hoping to find that missing cash.
Then, a thought crosses your mind, “Did I get the right change at the store?” Or maybe, just maybe, “Did someone sneak a bill or two?”
As frustrating as it is to misplace a bit of cash, imagine the scale when it comes to government funds. The amount of funds lost by governments each year is truly staggering. The Association of Certified Fraud Examiners(ACFE) estimates that 5% of organizations’ revenue is lost to fraud each year.
As AI evolves, so do illicit activities against governments. AI in the hands of bad actors makes it easier for them to commit crimes because it helps them evolve their tactics. For instance, fraud and cybercrimes continue to grow at an alarming rate, siphoning a large portion of government funds from its coffers. But fraud and cybercrime investigators can use AI to outsmart scammers and crooks.
Outpaced by growing threats and limited resources
The rise in fraud and cybercrimes has outpaced the growth in staffing and resources dedicated to combating these threats. According to Thomson Reuter’s 2024 Government Fraud, Waste and Abuse report, almost half of survey respondents said their most pressing issue is the increasing volume of work they are expected to handle without additional investigators. Job vacancies and budget constraints, common challenges throughout government, have hit investigation offices especially hard.
Those factors alone are difficult, but they seem insurmountable for governments that use traditional methods to detect and prevent fraud.
Carl Hammersburg, senior manager of government and health care risk and fraud at SAS, says, “Agencies are turning their focus from finding the risks more quickly to focusing audits and investigations to reduce the time to completion and gain maximum return on time invested.”
Reuter’s report also found that the top methods for detecting fraud are still cross-referencing public record databases and using whistleblower tips. Interestingly, one in five respondents also said their departments are either using or planning to use machine learning and AI to detect suspicious behavior and anomalous billing patterns.
It’s promising that governments are starting to examine new technologies. But the downside is that while they are considering AI alongside their tech infrastructure and data capabilities, criminals are gaining ground with generative AI (GenAI), using it to create fake documents and synthetic IDs.
Fraud experts say: Step up with data, analytics and AI
SAS is well-known for its industry experts, who have conducted investigations and been fighting fraud on the frontline. I asked a couple of my colleagues to share their thoughts on what’s ahead.
“Generative AI is lowering the bar of entry for fraudsters in all aspects of life. From health care to social benefits and procurement, the ease and volume of fraudulent activity will dramatically increase. Governments and industries will have to step up their game in using the same technologies to handle these increases," said Tom Wriggins, senior manager of risk, fraud and compliance solutions
Governments need to turn to AI to outsmart malicious AI
Alena Tsishchanka, insurance practice lead
2025: The year of AI and GenAI
Copy: As governments embark on using analytics and AI to combat fraud, it will be essential to apply them to meaningful tasks that will have the greatest impact. Machine Learning models can analyze vast amounts of data quickly and accurately, identifying hidden patterns and anomalies. Tasks such as scouring documents that require hours of investigation can be performed with AI in much less time.
While identifying fraud risks is crucial, it’s impractical to give investigators more tasks than they can complete. So, rather than using analytics to simply surface suspicious trends, governments need to deploy analytics and AI to reduce the amount of time each investigator spends on a case.
Using thoughtful data and AI can be a force multiplier for staff and offer tremendous value to fraud examiners. Using AI to investigate crimes does not remove the fraud examiner from the process; instead, it makes their job easier and faster, allowing them to focus on more meaningful aspects of the investigation. By automating routine tasks and improving detection accuracy, AI can help reduce the time examiners spend resolving a case, enabling them to manage their workload more effectively.
