When the pandemic prompted a massive expansion of government benefits, just about everyone was caught off guard. Agencies at the Federal, state, and local levels scrambled to identify eligible people and get money out the door. Long-term underinvestment in benefit systems resulted in a great deal of confusion, long wait times for assistance, and systems that were simply unable to cope with operation at such a massive scale.
All of this bureaucratic upheaval created an ideal situation for fraudsters. Not long after the benefits started to flow, stories about misallocated and stolen funds began to appear. Some of these were isolated incidents, where ineligible people found unique ways to tap into the flow of government assistance. In parallel to these small-scale incidents, a much larger (and more concerning) pattern of systematic fraud also emerged, where organized rings of criminals received benefits on a much larger scale.
Changing Patterns of Fraud
Before the pandemic, government agencies used employers to establish the legitimacy of claims, making the detection of fraudulent activity a relatively simple proposition. Using established patterns generated by decades of benefit distribution, they had a strong sense of what fraud looks like and how to prevent it. Though benefits fraud was always present, it wasn’t happening at the same scale.
When legislators expanded eligibility, they severed the link between benefits and employment, eliminating the ability to confirm identities with employers in the process. Add to that the hastily built nature of new benefit programs, which were defined by legislators and translated into payments in a matter of weeks, and the system becomes ripe for fraud. There simply wasn’t enough time to apply robust protections – the priority was getting help to people who needed it, fast.
Fraud prevention systems also failed to adapt to changing patterns of malicious behavior. Just as the typical persona of a government beneficiary changed dramatically over the past year, so did the persona of a typical fraudulent actor. Behavior which looked highly suspect before the pandemic became normal, and vice versa. Systems used to detect and prevent fraud weren’t adjusted in time to catch the “new normal” of deception in time.
The pandemic may be waning, but the new patterns of fraud it left behind remain an enormous challenge. Government agencies, auditors, and oversight bodies need a more adaptable, resilient system which can detect and prevent fraud as it changes before our eyes. That means not only learning from historical data, but taking action to predict and address the indicators of fraud as they emerge.
New Detection Tools
Artificial intelligence, machine learning, and adaptive cloud infrastructure have a strong role to play in identifying these signals and building the tools agencies need to deal with the new types of fraud that emerged over the past year. Given the huge amounts of data involved, combing through all of the potential indicators and pinpointing anomalies will require a great deal of sophisticated data analysis, not to mention a lot of focused computing power.
Here at Deep Labs, we’ve developed a rich set of models and tools which leverage the power of artificial intelligence to detect fraudulent activity in large data sets. We do this by creating personas which can distinguish between normal, legitimate payment activity and anomalous behavior. By applying these models to payments and benefit eligibility decisions, government agencies will be able to identify malicious activity with surgical precision.
To apply those models with the performance, scale, elasticity, and security government agencies require, we recently announced a partnership with Snowflake. Pairing the Snowflake Data Cloud with Persona-Based Intelligence from Deep Labs allows government agencies to extract powerful insights about fraud, waste, abuse, and mismanagement from large data sets. It also gives government agencies the flexibility they need to manage data using a single and seamless experience across multiple public, FedRAMP authorized clouds.
This powerful partnership gives government agencies the adaptable, resilient, scalable tools they need to combat changing patterns of benefit fraud. As we emerge from the pandemic with a changed sense of what’s “normal” in the world of payments, government agencies will need every tool in their arsenal to stay one step ahead.
Learn more about the partnership between Deep Labs and Snowflake.