The rise of AI insurance fraud threatens both the industry and consumers. We need to work together to combat this growing threat.
Across the world, the emergence of AI and deepfake technology has upped the ante in the fight against insurance fraud. AI generated images are being created in droves to trick insurers into processing claims when the items are, in fact, in perfect condition.
From photos of cars doctored to look like they have been in serious accidents to household contents with realistic looking damage plastered on top, the arms race between insurers and fraudsters has escalated.
Funeka Ngewu, Executive Head of Claims Operations and Support from Momentum Insure says the issue is on the rise in the South African market as well. “This new technology has drastic consequences for the industry as well as honest consumers. An increase in fraud risk only means claims processing costs are going to rise as we can no longer fully rely on traditional automated claims systems.”
Ngewu says these advancements necessitate the development of equally sophisticated AI-driven detection systems. These systems will need to be designed with agility in mind as the advancement in AI and deepfake technology often outpaces the industry’s ability to combat them.
Growing threat
The use of AI and deepfake technology in short-term insurance fraud is a relatively recent phenomenon, but it has quickly evolved to become a significant concern for the industry.
In the early 2010s, insurers began adopting AI for tasks such as risk assessment, customer service, and claims processing. “AI’s ability to analyse vast amounts of data quickly and accurately made it a valuable tool for improving efficiency and reducing costs,” says Ngewu.
By the mid-2010s, major advancements in deep learning, a subset of AI, led to the creation of deepfake technology. Originally developed for entertainment and research purposes, deepfakes use neural networks to create highly realistic but fake images and videos.
By the late 2010s, fraudsters began exploiting deepfake technology to create convincing images and videos to support fraudulent insurance claims.
Industry response
As deepfake technology became more sophisticated, so did the methods used by fraudsters. Insurers responded by investing in AI-driven detection systems designed to identify deepfakes and other fraudulent activities.
These systems analyse image metadata, check for inconsistencies, and use machine learning to spot anomalies. Today, Ngewu says the battle between fraudsters and insurers continues, with both sides constantly refining their techniques.
“As an industry, we have no choice but to focus on collaboration, sharing data and insights to strengthen defences against this evolving threat. However, as the claims process becomes bloated, the cost increases,” warns Ngewu.
She says the situation is made worse as syndicates have fully leaned into AI technology as a means to profit from insurers. “These syndicates have been defrauding insurers for decades and all they end up doing is pushing premiums up for law-abiding policyholders. It’s a continuous cycle that affects everyone involved, from the insurers trying to mitigate fraud to the honest consumers who bear the brunt of increased premiums.”
Ngewu warns against the continued rise in AI-generated insurance fraud. “We are always in the process of upgrading our defences, investing in similar AI-driven technology to detect fraud faster.”
However, she says this becomes increasingly difficult when AI technology experiences an exponential rise in complexity and accessibility. “The industry and all its stakeholders need to remain vigilant in these tech-driven times. The insurance industry is constantly evolving,” Ngewu concludes.