How Technological Advances are Revolutionizing Insurance Fraud Detection

Fraudsters beware: technological advances mean insurers will catch you As alarming as these kinds of scams are, they will become increasingly difficult to pull off.  Advertisement As alarming as these kinds of scams are, they will become increasingly difficult to pull off.

Devotees of true crime documentaries and podcasts will be all too familiar with the trope of spouses taking out life insurance on their partners before murdering them for the payout. While it’s not possible to do that in South Africa without the party’s consent, that doesn’t mean we don't have our own rogue’s gallery of insurance fraudsters. DStv and Showmax subscribers, for example, may be familiar with Rosemary’s Family. 

The documentary tells the story of police officer turned serial killer, Rosemary Ndlovu, who murdered partners and family members after taking out funeral policies on them. More recently, an alarming story emerged of a prisoner and his wife making claims on lost items such as cellphones, laptops, and video cameras. By keeping the value of claimed items under R10 000, the prisoner avoided triggering an assessor’s visit and successfully defrauded insurers until he was finally caught. The real kicker? He was already serving 12 years in prison for defrauding Mutual and Federal Insurance in 2016.

As alarming as these kinds of scams are, they will become increasingly difficult to pull off. Even though incidences of fraudulent and dishonest claims have risen in recent years and scammers have become more sophisticated, the industry is making rapid gains in ensuring fraudsters can’t get away with committing the crimes they do. How fraudsters operate According to Mishaya Chettiar, Executive Head at Everything.Insure, a big part of tackling fraudsters is knowing how they work and the techniques they use.

More on this Trio who allegedly forced admin clerk to perform fraudulent transactions worth over R8 million, in court Standard Bank says measures in place to protect customers after flurry of fraud complaints Bank employees sentenced for money laundering “Because the insurance sector is so large, diverse and complex, it is also unfortunately highly susceptible to financial crime,” she says. “All it takes is for someone to record a few incorrect details on a claim for fraud to occur.” Further complicating the situation is that new technologies, such as generative AI, have made it easier to commit fraud and to do so at scale.

 “Generative AI means that it’s possible to create things that did not exist, such as photoshopped and AI images with fraudulent backgrounds,” says Chettiar. “If the insurer does not have the tech to detect that these are AI-generated images or documents, it will pay out fraudulent claims.” “Clients could also create fraudulent insurance documents pretending that they have taken out insurance on something they don’t even own, pay for a few months and then wrongly claim,” she adds. “Ultimately, digital crimes are much easier to perpetrate than in-person ones.” Not a victimless crime Another factor to bear in mind when it comes to insurance fraud, as Chettiar points out, is that many perpetrators view it as a victimless crime.

“Many claimants see their actions as a victimless crime, only impacting ‘wealthy insurance companies with deep pockets’”, she says. “But they’re not. Fraudulent claims, even if unintentional, impact every single policyholder, no matter the size of the claim.” As Chettiar points out, falsified claims or exaggerated losses drive up overall insurance costs for everyone, further impacting an industry under severe strain. In some cases, insurance premiums have increased by 40 percent or more due to several factors, such as Covid and falsified claims.

Chettiar adds that some people also feel that they’re entitled to make fraudulent claims because they’ve paid into a policy for years without making any claims. But no matter how much you pay into a policy, chances are you’ll never match the amount of the item you’re insuring. This is especially true for big-ticket items such as cars. As Chettiar notes, if you have a R500 000 car insured at R500 a month, it would take 83 years for you to cover the cost of the insurer replacing the car. 

“Insurance is a communal offering,” she says. “In order for it to work effectively, it’s vital that people play by the rules and don’t try to jimmy the system.” Fighting back Though technology is certainly playing a role in making it easier to falsify claims in a way that seems legitimate, it is also key to helping the insurance industry combat fraudulent crime.

Tools like AI and machine learning are helping greatly, to identify patterns and anomalies within vast data sets, enabling insurers to detect fraudulent claims with greater accuracy and efficiency. And, because AI algorithms can continuously learn and evolve through new data, insurers can adapt to new fraudulent tactics, keeping pace with any fraudulent actors. It is largely due to technological advancements and the work done by organisations such as the National Insurance Crime Bureau (NICB) and the South African Insurance Association that the industry is gaining significant traction in its fight against insurance fraud which is evident in the latest fraudulent claims stats from ASISA. While the number of fraudulent and dishonest claims rose substantially in 2022, the losses felt by these claims were only R77 million instead of the R1.1 billion it would have been had it not been for the use of big data, machine learning, AI, improved data sharing, and enhanced authentication mechanisms such as biometric customer identification. 

“Detecting fraud can be a costly affair and is a process that takes a long time and requires vast resources to complete,” adds Chettiar. “AI, machine learning, and other technologies are helping to detect fraud much faster and easier by enhancing risk assessments and providing more insight into insurance applicants by combing through greater data sets. More than this, AI can be used to identify and flag applications that have been created using other AI tools to commit fraud.”