Advancements in technology are making insurance fraud easier to commit and harder to detect.
Digital fraudulent activity happens on an expansive scale, from one-off manipulated insurance documents, right up to global organised crime rings, feeding into larger crime activity such as illegal substances and people trafficking.
Just as criminals are upping their game by using tech and AI to commit fraud, we’re upping our game too. We’re harnessing the power of artificial intelligence and combining it with years of human intelligence, experience and expertise to tackle the rise in digital insurance fraud.
Let’s explore the rise of shallowfakes, emerging industry trends and the impact they’re having.
What is a shallowfake?
Shallowfakes first appeared around seven years ago but became rapidly prevalent in the post-pandemic world. Fraudsters quickly recognised that more and more insurance companies were digitising the handling of claims. In the past, documents would have been forged requiring a physical action, like copying a person’s signature. With shallowfakes, documents are altered digitally.
The purpose of them is usually to mislead people about someone’s identity or to make false or exaggerated insurance claims.
As the name suggests, they’re a less technical version of their partner in crime, deepfakes. Whereas deepfakes will include sophisticated AI and machine learning, shallowfakes are digitally manipulated documents, with changes to text, data and details. However, despite their simplicity, they can be more effective, easier to implement and difficult to counter.
Shallowfakes began to appear in various guises to enable people to set up bogus insurance policies and bank accounts or invoices to support claims. Instances of organised fraud opened the door for opportunists when it became clear that shallowfakes are relatively easy to produce.
Pretty much anyone has the potential to produce a shallowfake. This might be a professional scammer making up insurance claims to get big payouts, or just an ordinary person looking to earn some extra money by overstating their claim.
They’re easy to make, they’re easy to replicate and they’re hard to stop.
The impact of shallowfakes
- They cost insurance firms serious money in fraudulent claim payouts
- They fuel rising insurance premium costs
- They cost time and money for claims teams in an insurance firm
- They have the potential to make a firm a target for fraud rings
- They can cause serious damage to a firm’s reputation
- Customers grow concerns around a firm’s data security
- As a result, they drive up costs for the consumer, which could result in either dissuading people from taking out insurance policies in the future…
- …Or taking out cheaper policies with less reputable providers
How we linked 13 false claims
We recently carried out document validation on a possible fraudulent claim. What we discovered was a huge amount of information, tying this claim to 13 others that our client was handling.
The deeper we went, the more we found. Every claim was false and all of them were using shallowfakes to manipulate documents like medical records in very convincing ways.
The estimated value was in the region of £80,000, but it didn’t stop there. We identified that the culprit had hit different insurers using the same shallowfake documents and names, with similar high value claims. We tracked the person behind it to another country, and the fraud was probably worth hundreds of thousands of pounds.
Now trending – what patterns do we need to be aware of?
Fraudsters are using a combination of shallowfakes and AI (deepfakes), helping them keep one step ahead of detection. Most claims are for specialised products like bicycles, travel or collision waivers, where insurance companies rely on digital processes. For these insurers, the cost of thorough investigations often outweighs the benefits.
Vulnerable groups are especially targeted. Fraudsters take advantage of these ‘clean skins’ – individuals who have no prior record of insurance claims or fraud – making them appear more trustworthy and low-risk to insurers. They are exploited due to a lack of prior suspicious activity, so the likelihood of the fraud being detected is reduced.
The issue and its impact is growing. Fraudsters are being identified as operating on an international scale, submitting claims to several insurance companies using deepfakes (AI) and shallowfakes. The growing number of bogus claims is taking a longer and longer time to resolve.
Where AI is being used by some of the larger companies to combat deepfake claims, less attention is being directed to the rapid increase of shallowfakes in the industry – they are so easily overlooked.
The numbers you need to know
- One in every four documents show some sign of tampering
- Just under one in ten of these contain high fraud risk indicators, indicating they are likely fraudulent
- That means that claims processors could automate between 75-92% of their document validation…
- … which could increase their processing capabilities by up to ten times the efficiency
- In 2024, a 300% increase in shallowfake fraud was reported, which has led to a steadily rising cost of insurance premiums since 2022, increasing by £200 on average
- And the average cost for insurers?
- Life insurance: £73,578 per claim
- Home insurance: £5,284 per claim
- Car insurance: £3,479 per claim
The rise of AI
AI may only play a small part in creating shallowfakes, but a big part in combatting them. Basic editing software and a user with a bit of skill can produce high quality shallowfake documents easily and quickly. AI finds these accurately, quickly picking out the falsified documents from the real ones.
Detection methods are evolving. AI is equipping us with new fraud detection models that can free up our investigators to do what they do best – follow up on individual cases and apply context, empathy, experience and skill.
Keeping business safe through Lynx Documentation Validation
Shallowfake activity is costing the insurance industry and its clients a huge amount of money. Something needs to change. Detecting, challenging and defending these kinds of fraudulent activities is essential.
Using the latest AI software in Lynx Documentation Validation, combined with years of knowledge, experience and true human intelligence, we’re leading the charge against digital insurance fraud. We work with you to develop our services especially to tackle the challenges of this ever-evolving problem.
Ready to safeguard your business against costly shallowfakes?