Preventing Benefits Fraud: How AI Helped NY State Protect Public Funds

Key Challenges

The New York State Department of Labor (DOL) had to overcome a staggering increase in fraud attempts and the limitations of manual processes in handling large data volumes and AI-driven fraud tactics, while meeting increased demand for benefits and protecting billions in public funds.

Key Results

With Splunk, the New York State DOL proactively identified over 1.5 million fraudulent claims and prevented more than $32 billion in attempted theft, improving the speed of investigations and overall fraud risk management.

Download PDF

When demand for government benefits skyrocketed¹ during the pandemic, this led to a new wave of fraud.

Bad actors stole billions of dollars from benefits programs, according to estimates from the Government Accountability Office.²

In New York, the DOL experienced a staggering increase in fraud attempts during the pandemic. Fraud attempts jumped from 94 in 2019 to more than 400,000 by 2020 and more than one million attempts by 2021.³ During this same period, the department paid out more than $105 billion in benefits, equating to 50 years’ worth of benefits.

DOL’s situation underscores the dual challenge state and local agencies face: meeting increased demand for assistance while keeping fraudsters at bay. Artificial intelligence (AI) and data analytics can help agencies mitigate fraud and bolster program integrity, ensuring that only eligible individuals receive benefits. “There’s a cost to fraud,” says Tina Carkhuff, a government industry advisor at Splunk. “There’s a cost to all of us as individuals. There’s a cost to our federal government and our state governments. We don’t want to lose money that people are entitled to get.”

Outcomes

1.5 million fraudulent claims identified
$32 billion in attempted theft of benefits prevented
AI fraud detection

From manual to AI-powered fraud detection

Like many agencies, DOL relied on manual processes for parts of its operations. The agency used spreadsheets to review data and identify patterns that indicated potential fraud, such as multiple claims across different addresses tied to the same bank account.

The pandemic showed that manual approaches don’t scale. Fraudsters now use AI in their schemes, bringing a level of automation to cybercrimes that’s difficult for organizations to combat with human intervention alone.

It’s why agencies are turning to this same technology to strengthen fraud mitigation and prevention. AI and data analytics can easily surface patterns and anomalies in large volumes of data, allowing employees to pinpoint suspicious activity in minutes rather than days or weeks

DOL adopted the Splunk App for Fraud Analytics for this very reason. The solution includes fraud analytics as well as monitoring and detection tools. It seamlessly receives data directly from the agency’s legacy system. The department has used the app to complement the work of its Office of Special Investigations and to assist in highlighting patterns of suspicious claims for further review by human investigators.

Harnessing the power of AI and data analytics

New York is a populous state, so DOL data sets can be incredibly large. As opposed to teams reviewing huge spreadsheets, Splunk can filter the data and allow humans to make qualitative judgments more quickly. This streamlines repetitive tasks, giving teams more time to focus on higher-value tasks.

State and local governments can stay one step ahead of cybercriminals with the right technology. Since the pandemic and before implementing Splunk, the special investigations team of NY DOL has:

Identified more than 1.5 million fraudulent claims

Prevented more than $33 billion in attempted theft of benefits

AI and data analytics help agencies modernize claims processing, improve fraud risk management while abiding by various regulatory compliance requirements. These tools also ensure legitimate claims are processed in a timely, accurate manner because they can reduce the time employees spend manually investigating fraud.

DOL began to reap these benefits when it worked with Splunk’s professional services team throughout the implementation process in 2023. Specifically, Splunk worked with the New York State Office of Information Technology Services to build out Splunk Enterprise Security for DOL to leverage the Splunk App for Fraud Analytics.

DOL has used Splunk’s reports and data dashboard to isolate different bank accounts, overcoming the data standardization and formatting challenges.

With the Splunk dashboard, DOL investigators now access updated data faster dramatically saving reporting run times. The department also plans to use Splunk’s solutions to mitigate phishing attacks against legitimate benefits claimants. During the pandemic, fraudsters used more than 1.5 million stolen identities. DOL wants to use Splunk’s solutions to make sure cybercriminals can’t use these stolen identities to apply for benefits.

Conclusion

The pandemic gave fraudsters new and even bigger targets, making them more relentless than ever. As criminals increasingly use AI in their schemes, it’s critical for governments to leverage AI to prevent widespread fraud across their ecosystems. With AI and data analytics, agencies can be equally unrelenting in defending benefits funds and public trust.

There’s a cost to fraud. There’s a cost to all of us as individuals. There’s a cost to our federal government and our state governments. We don’t want to lose money that people are entitled to get.

Tina Carkhuff, Industry Advisor, Public Sector, Splunk