The 13th annual RGA Fraud Conference highlighted emerging themes in the fight against fraud, including the use of technology as both a weapon of deceit and a line of defense.
An exterior photo of a palatial home in Staten Island, NY, features finely manicured landscaping, a brick-inlayed driveway and a front staircase edged by two imposing gargoyles.
In an area not known for its affordable cost of living, the home – purchased entirely with the proceeds of a well-organized life insurance fraud scheme – served as an important reminder for attendees of the 13th annual RGA Fraud Conference: Insurance fraud can be very lucrative for criminals and very costly for insurers.
The three-day online event in August provided insurance leaders with deeper insights into the tools and tactics fraudsters are wielding, as well as the lates statistics and innovations to battle back.
Here are some highlights from the conference sessions:
Machine learning and deep learning can be trained to detect anomalies in data that may indicate fraud. AI’s ability to process vast amounts of data and recognize patterns makes it well-suited to detect complex fraud schemes such as rebating, stacking, churning, ghostwriting, and account takeovers. But human intelligence remains essential to interpret these anomalies and validate red flags. Understanding the mindset and tactics of fraudsters is essential. Carriers must continually adapt their defenses by using, yes, AI, but also behavioral science and cross-functional human collaboration.
“Artificial intelligence is just modeled after human intelligence. The machine can learn, think, and reason like a human can — to a degree."
Taking a siloed approach to fraud prevention is fraught with peril. Instead, using a lifecycle approach can help insurers better understand the customer and how a change in behavior might suggest an attempt at fraud, including potential identity fraud.
Invest in technology across the policy lifecycle to detect and prevent fraud; break down internal silos.
“We are in an AI arms race, and the truth is that fraudsters are moving much faster than we are.”
The US life insurance industry paid out $89 billion to beneficiaries in 2023, a total that nearly matches what the US pays in Social Security benefits annually. In doing so, fraud costs each life insurance carrier an average of $300 million each year.
Partnering with law enforcement is vital to fight fraud. Merely terminating bad agents or rescinding policies when fraud is discovered allows bad actors to pop up again in a different place to do the same thing again and create more victims. The spill-over effect raises prices for those who are law abiding and honest and reduces access to the financial protection that insurance offers.
“This is where the noble work really comes in. We talk in our industry about a 40% coverage gap, the percentage of individuals who have either not purchased life insurance or don't have adequate life insurance. I know life insurance is relatively inexpensive, but cost is still the primary reason. It's our responsibility to bring that cost down so that every individual can afford life insurance.”
Rationalization enables dishonesty. When people believe they are entitled to something, they can justify doing something wrong, such as being dishonest on an insurance claim or application. They are then able to do something that they know is wrong without creating any self-harm or guilt. If insurance companies normalize dishonest customer behavior, they assist in creating an environment where people will be dishonest.
Also, fraud risk is not just from those people who set out to be deliberately dishonest. It also comes from otherwise good people who, in certain situations, will choose not to be honest. Insurance companies need to recognize this and be aware that people will not always be honest.
When reviewing steps in our processes where we look at automation, recognize that if we can humanize the insurance process, we create an environment where people are more likely to be honest. Look for ways to use messages to promote positive behaviors. For example, focus on the 95% of customers making honest disclosures rather than the 5% who are not.
“We don’t have barriers; we have honest customers. This concept has the potential to be applied to how insurance companies approach both application and claim forms. It is a reasonable positive statement that endorses the right behaviors.”
The threat of benefits fraud and abuse in the Canadian health insurance industry is growing, but collaborative efforts are transforming fraud detection and prevention.
Industry-wide initiatives such as data pooling, joint provider fraud investigations, and educational campaigns are key to fighting fraud. These tools enable insurers to identify suspicious patterns, share evidence securely, and take coordinated action against fraudulent providers. Case studies illustrate the effectiveness of these programs, demonstrating significant financial savings and improved fraud mitigation through joint investigations and AI-powered analytics.
“Benefit fraud threatens the sustainability of our plans, their affordability, and access for the 27 million Canadians who rely on their benefits. It’s not a victimless crime. It affects us all.”
“There’s a lot of money at stake, and the fraudsters know this. Now it’s globalized, so it can be committed from virtually anywhere, which can make it hard to detect and prosecute.”
Generative AI improves data analysis, uncovering complex patterns that allow for better decisions and improved fraud detection. AI can learn from new data over time, but that data still should be checked by humans to guard against AI “hallucinations,” where AI completes gaps with what it thinks should be there but, in truth, is fictional.
Incorporate AI into fraud detection, but balance that technological push with the skills of talented humans to verify AI’s output.
“Machines don’t get bored. They approach situations with the same level of commitment and excitement from the first time to the 1,000th time.”
“Illusion and confusion. Fraudsters deliberately create illusions — altered digits, false addresses, fake labs — to cause confusion, betting that insurers will give up and pay out."
Save time, fight fraud, and improve claims outcomes with RASCL, RGA's automated solution for claims leakage. Learn more today.