Insurance companies are cracking down on scams that employ the use of artificial intelligence (AI) in South Africa.
According to Ayanda Mdhluli, a senior manager of cybersecurity at Old Mutual Insure, the group has seen an alarming increase in the use of AI to scam people or conduct insurance fraud.
Most commonly, these individuals will attempt to use AI to fabricate or manipulate the documents needed for an insurance claim.
For example, a perpetrator may use AI to create a false image of their car with a damaged door, and then attempt to submit the image as visual evidence to support a claim that they were involved in an accident.
In response, insurance companies are ramping up their validation techniques, employing their own AI tools while also relying on external sources to back up a claim.
“We are also embracing AI as a tool to protect our customers, improve efficiency, and enhance fraud detection,” said Mdhluli.
The end result is that insurance companies are adopting more stringent assessment criteria for claims and can no longer rely purely on photo evidence.
In Old Mutual’s case, the group collects evidence from additional first-hand sources that were at the scene of an alleged accident, such as emergency responders and tow truck drivers.
“Relying solely on images – especially as AI-generated visuals become more sophisticated – is no longer sufficient,” it said.
It also uses AI tools that allow the company to detect whether an image was artificially generated, debunking the claim in the process.
“We take this very seriously. When we identify a fraudulent claim, we report it to the Insurance Crime Bureau, which alerts the wider industry,” stated Mdhluli.
“There are consequences for this type of behaviour, including blacklisting, a cancellation of your insurance policy, and even jail time.”
Various insurance providers are also ramping up their use of tools like geo-tagging, biometrics, and geometrics to discern when a person’s story doesn’t add up.
A growing problem
Mdhluli explained that, while generative AI is still fairly new, there have already been plenty of scams targeted at companies and consumers alike.
Old Mutual has experienced cases where the same photo of a damaged car were used in multiple claims.
At other times, perpetrators have simply downloaded images from the internet to support fabricated invoices.
Unfortunately, these tactics are becoming harder to identify as a result of AI, which can conjure a convincing image of a damaged vehicle from scratch.
Old Mutual added that they have started to receive claims from multiple sources, including vehicle repair locations.
“We’ve picked up repeated invoices with headers from third-party suppliers like glass fitment centres or repairers and are starting to see the same invoice submitted across different claims,” it said.
In one instance, an individual staged an accident, dropped off a vehicle at a panel beater, and was linked to a syndicate operating across provinces.
These kinds of insurance scams are on the rise in South Africa, he said, particularly in economically strained regions like the Eastern Cape, according to BusinessTech.
The issue extends to the pre-owned market, too, where unscrupulous individuals fake important information about cars before selling them to buyers who haven’t exercised due diligence.
In these scenarios, the onus is on the policy holder to ensure that all of the information they provide to their insurer is correct.
It is recommended that buyers ask for the car’s NaTIS certificate (the official registration document validating the legal owner) before proceeding with the purchase.
It’s a clear red flag when the alleged owner cannot produce the certificate, and buyers should also check the document when they receive it to make sure that everything matches.
Tell-tale signs of forgery include typos, inconsistencies, and incorrect fonts and formatting.
Loading ...
