AI-driven insurance fraud – from deepfakes to scams

0
130
AI-driven insurance fraud

Recently, I came across a viral AI-driven insurance fraud LinkedIn post that resonated with me. A woman faked her own car accident using AI-generated images. She created photos of a “smashed” car, added bruises to her own face, staged a crash scene and even produced a realistic-looking receipt.

The woman was not trying to file a real claim. Her goal was to highlight how hard it is becoming for the insurance industry to distinguish real claims from fakes. Sadly, this is not just a once-off incident. It points to a global challenge we’re now facing as insurers.

From deepfakes to scams

Artificial Intelligence (AI) tools like ChatGPT are readily available, and anyone can use them to produce fake content. That reality is forcing us to rethink how we verify and process claims.

We recognise the double-edged sword of this technology. Yes, AI-driven insurance fraud is changing the game. But we are also harnessing it to protect our customers, boost efficiency and detect fraud more effectively.

Our current claims process does not rely on a single source of evidence. We verify submissions using multiple inputs. Tow truck drivers and other first responders provide immediate details from the scene. On top of that, our assessment centres play a vital role in evaluating damage.

Expert panel beaters and professional assessors are appointed to physically inspect vehicles. Relying solely on images is no longer enough, especially now that AI-generated visuals are becoming so convincing.

Seeing through digital deception

While AI-driven insurance fraud might seem like a recent trend, I have seen many examples of scams where people are either unaware they have been duped or they’re actively trying to trick the system.

Some of the scams we have detected include the reuse of identical damage photos across multiple claims. Interestingly, we have also flagged images pulled from the internet and paired with fake invoices.

In one pattern, we noticed scammers used the same invoice multiple times. These were often from third-party suppliers like glass fitment centres or repair shops. It is a small detail, but one that can reveal a lot when you are tracking repeat claims.

There was even a case where someone staged an accident, dropped the car off at a panel beater, and turned out to be part of a syndicate operating across provincial borders. These scams are happening more often, especially in places like the Eastern Cape, where financial pressures are high.

That is why we are taking action and investing in geotagging, biometrics and geometrics wherever possible. We are also collaborating with the Insurance Crime Bureau to flag suspicious activity early.

We have also started using AI tools ourselves, not just to detect fake images, but to identify manipulated documents.

When we do catch a fraudulent claim, we take it seriously. We report it to the Insurance Crime Bureau, and the broader industry is alerted. Consequences can include being blacklisted, having your policy cancelled and in some cases, facing jail time.

Staying safe from deepfakes and dodgy dealers

In today’s economy, more people are under financial strain. That makes it even more important to stay alert and do your homework, especially when something seems too good to be true. The second-hand car market is a hotspot for scams.

If you’re buying a vehicle, be cautious. Do proper checks before you sign anything. We’re seeing more cases where scammers exploit buyers looking for a good deal. When you take out insurance, it is your responsibility to make sure all the information you give us is accurate.

We rely on that information to assess risk and give you the right cover. You also have specific rights and responsibilities when you enter into a policy agreement. As AI makes it harder to tell what’s real and what’s not, we are staying one step ahead to protect you, our partners and the insurance system itself.

How to avoid deepfakes and scams

Here are a few of my tips to help you avoid scams and steer clear of insurance trouble when buying a second-hand car:

  • Ask for the NaTIS certificate

The NaTIS certificate proves legal ownership. If the vehicle is still under finance, the bank will hold this document until the loan is paid. If the seller cannot produce it or gives a vague reason, walk away.

  • Match the details

Check that the VIN, engine number, make and model on the registration documents match the actual car.

  • Spot fake documents

Typos, inconsistent fonts or odd formatting could mean the paperwork has been forged.

  • Check for signs of tampering

Damaged locks or ignition systems might suggest the car was stolen.

  • Avoid foreign-registered vehicles

Cars from outside South Africa or neighbouring countries are often difficult to insure.

  • Look out for grey imports

Features like fender mirrors, headlight wipers or labels in other languages might mean the car was illegally imported.

  • Get a microdot certificate

This helps confirm the car’s identity and will support your claim if something goes wrong.


Ayanda Mdhluli | Senior Manager | Cyber Security | Old Mutual Insure | mail me |





LEAVE A REPLY

Please enter your comment!
Please enter your name here