Weightmans partner and Head of Fraud Mike Brown reflects on how AI isn’t just enabling fraud on a greater scale, it’s beginning to normalise it, creating a challenge for insurers and businesses that extends far beyond technology and into consumer behaviour.
Cast your mind back – and be honest. We’ve all been there at some point in the dim and distant past. Sat in the nurse’s office at school, clutching ourselves and turning a mild twinge into a crippling pain. Presenting a playground graze of the knee as an agonising wound and pleading with the school secretary to call our parents to take us home.
Exaggerating our challenges isn’t anything new, and it’s probably something we’ve all done at some point. Unfortunately, exaggerating an insurance claim or a compensatory complaint isn’t new either – but the AI revolution threatens to throw up stunning deceptions that would fool even the most eagle-eyed school secretary.
The rise of generative AI has fundamentally changed the rules for insurers. Photographs, identity documents, accounts and invoices – even voice recordings – can now be easily invented or altered by AI programs that a fraudulent claimant doesn’t even need to pay for.
Aviva’s latest fraud report illustrates the scale of the challenge. Alongside detecting, well over £230m in fraudulent insurance claims last year alone, across nearly 20,000 suspect cases, the insurer warned that AI-generated images and manipulated documents are becoming increasingly common. As the company’s Head of Claims and Counter Fraud noted, AI was enabling the efforts of would-be fraudsters to become more sophisticated, making it increasingly complex – and costly – for insurance companies to detect such scams without driving up premiums and scrutiny for genuine consumers.
From obvious examples like imagery, where AI programs can convincingly generate damage to an intact windowpane, all the way to bogus documents, AI has a wealth of content to draw upon in its scope to invent a textbook deception. One particularly concerning trend is so-called “claims inflation”, where genuine losses are exaggerated using AI-generated imagery or documentation to secure larger payouts. Suddenly that genuinely loose or cracked windowpane becomes shattered.
The insurance industry can’t simply accept these developments as the new cost of doing business. Nor can it continually pass the financial burden of fraudulent claims onto honest policyholders through higher premiums or more intrusive scrutiny or legitimate claims. And yet if AI has changed the rules there is another concern that upends the entire game.
The more significant challenge isn’t simply that AI makes fraud easier. It’s that making deception so simple and accessible risks changing how consumers think about it. On the one hand, a sceptic could argue this is to be expected. Hand everyone access to a technology that circumvents established rules and norms around ownership, creativity and license, technology with the ability to pull documents and images seemingly out of thin air, and of course there will be a growing temptation to break the rules yourself. Particularly when consumers are being told that living standards are stagnant and price gouging is rife.
The danger is that what once felt like outright fraud begins to feel more like harmless optimisation. Editing an image, tweaking an invoice or exaggerating damage may no longer be viewed by some as criminal behaviour, but simply as making sure they receive what they believe they deserve.
The evidence suggests this cultural shift is already under way.
Global research from Verisk found that 36% of consumers would consider making a rule-breaking edit to support an insurance claim, while among Gen Z respondents that figure rises to 55%.
Insurers and anti-fraud professionals therefore face a gruelling uphill battle. An insurer attempting to navigate this fresh landscape alone is swimming against a rising tide – AI tools will continue to adapt at an alarming rate, and predatory fraudsters will pursue firms that are least prepared, peeling off the slowest runner at the back of the herd.
This makes collaboration more important than ever. In the UK, many firms are members of pre-existing industry bodies and enforcement partnerships to streamline their operations and boost efficiency; on the international stage there are networks of expert investigators that share counteroffensive strategies and monitor the emerging trends. Strengthening and building upon these relationships is imperative.
This is easier said than done. What we’re contemplating is a culture shift, but an essential one to protect the market and honest policyholders. The global nature of AI fraud also makes water-tight governance processes essential. The sharing of intelligence across jurisdictions, many of which will have different privacy and information sharing regulations can be incredibly complex. It is entirely possible for watertight cases of fraud to collapse if governance laws and processes in one country haven’t been adhered to in another.
Accommodating tech developments in industries like insurance was never going to be easy, and it was never to come without disruption. There is no question that AI is forcing firms to step out of long-held comfort zones and adapt to a rapidly changing environment. As fast as they can respond to AI developments, and as closely as they can collaborate in the fight against fraud, they must also be at the centre of a wider conversation about the tolerance of unethical, illicit behaviour, and be a key driving force in the push to encourage honesty from consumers.
Insurers aren’t simply playing to win against AI-enabled fraud. They’re facing an evolving threat from fraud being increasingly seen as normal behaviour. And if the debate is lost on that front, then the game as we know it might just be up.
Be the first to comment