Fraud has always been a problem for the insurance industry, but as the threat evolves and scams become increasingly high-tech, insurers are being urged to fight fire with fire.

The figures around insurance fraud are eye-watering: Insurance Council of Australia (ICA) members estimate undetected fraudulent claims cost the Australian insurance industry around $400 million in 2023. These undetected cases drive up insurance premiums and result in significant harm to customers and the economy, according to the peak body.

In 2022, the US National Insurance Crime Bureau said general insurers had paid up to $US9.2 billion extra in disaster claims the preceding year due to scams. ICA – which was to hold a session on fraud prevention at its annual conference on October 16 – has since appointed former cybercrime squad detective Andrew Gill to lead a standalone scam-fighting unit focusing on organised fraud networks and emerging trends.

The problem, as ICA has noted, is being compounded by advanced – and ever-advancing – technologies. Allianz recently warned “shallowfake” scams could be the next big threat to the industry, with fraudsters using relatively simple tools such as Photoshop to edit “damage” into vehicle photos submitted in claims. Cases involving the distortion of real images, videos and documents have tripled in number recently, the insurer says.

And at the more high-tech end of the spectrum, generative artificial intelligence (GenAI) threatens to take fraud to new heights.

Insurers must guard against sophisticated rorts that can range from deep faked claims to automated, large-scale submissions of false claims. Earlier this year, ICA told a Senate committee that GenAI could be used to augment photos, and will help criminals with phishing, voice spoofing and identity fraud. 

Drew Whitmore, head of global partnerships at insurance AI specialist Shift Technology, says scammers are already tapping into GenAI. “We’re seeing examples like individuals using ChatGPT to write an accident script,” he explains. “When they call in to report the claim, they have written a perfect accident description that is hard to detect as a potentially fraudulent accident.” Mr Whitmore says fraudsters submit claims 58% more frequently than other individuals, and many are marked as high severity. While claim severity has increased across the industry, fraudulent claims show even more dramatic trends. “When analysing the claims of bad actors committing fraud, the severity of those claims is over eight times more than the standard,” he says.

New technology allows fraudsters to operate with precision and scale. Social media has also helped them, providing a platform to gather personal information and craft convincing claims.

Fraudsters exploit publicly available posts to fabricate accidents, for example, making it challenging for insurers to verify the legitimacy of claims. Abhi Bhola, growth leader for APAC at data, analytics, and insurance operations company EXL, says the scam threat is growing. “Particularly things like deepfakes, that’s scary,” he says. “We see so much information available for a malicious mind to work through. “We are seeing a lot of new approaches getting into the fraud process, and I think that will only be going higher with better AI, which is easily available.”

But if AI is the problem, it can also be the solution: insurance companies can harness it to fight back against criminals.

AI tools allow insurers to sift through large volumes of information in real-time to identify anomalies and flag potentially fraudulent claims before they escalate. EXL and Shift Technology have teamed up to help the industry do just that, combining EXL’s data, analytics, and insurance domain expertise with Shift’s AI-driven fraud detection solutions.

Mr Bhola says: “Undetected fraud has a real cost for insurers and policyholders alike. We aim to offer solutions that directly address this by enhancing fraud detection accuracy.” He says the partnership gives insurers “a comprehensive solution to improve proactive fraud detection and operational efficiency”.

He notes that insurers are expected to contribute to society by ensuring fair treatment and protecting customers from exploitation.

“Insurers want to be ahead of the game. It’s not just about reducing fraud losses – it’s about demonstrating to policyholders that their insurer is doing the right thing, ensuring fair treatment, and fulfilling their role within the community.”

All this requires a cultural shift, according to Mr Whitmore. “We must move from reactive to proactive,” he says. “If insurers cannot do that, then leveraging any advanced technology will not help.”

Indeed, as fraudsters evolve, so must the industry – and Mr Whitmore hopes regulators will follow suit, particularly on the issue of data sharing.

 “I think as we go along into the role of AI and extensive use of data, we’ll see a bit more collaboration from the regulator, which will understand this importance to stay ahead of the game,” he says. “Fraudsters are sharing their information, and the industry needs a framework to do this, too.” 

“Our global experience shows that leveraging both internal and external data, along with analysing network-level linkages, is essential to staying competitive against industry threats,” Mr Bhola adds. “By enhancing insights through insurer and publicly available data and utilising advanced AI for fraud detection, we can build the foundational capabilities needed to uncover organised fraudulent network activity.”

Mr Bhola sums up the required approach: “Combining data, leveraging AI and building collaborative ecosystems and partnerships across the insurance industry is what it takes to enable a proactive fraud management regime and stay one step ahead of fraudsters.”