The AI arms race: criminals are sharing data. Why aren’t we?

By Simon Bednar, Finsure Group CEO

There is a war raging in our industry right now, and lenders, aggregators and brokers are on the same side. At its heart is AI, and on the other side are criminal networks leveraging new technologies with new scale, speed and sophistication to exploit weaknesses across the lending ecosystem.

Scams and fraud used to be limited by the bandwidth of criminals. It was a one-to-one business. Now, an AI agent can run 1,000 different scams to rapidly target vulnerable people or institutional vulnerabilities simultaneously.

AI is driving efficiency and creating value for our industry, but on compliance, the AI honeymoon is over as we are starting to see its capabilities in the world of mortgage fraud. Nefarious individuals are using AI to create realistic but fraudulent documentation that can bypass robust systems and checks.

As fraud becomes increasingly sophisticated and organised, lenders, aggregators, brokers and regulators need a whole-of-industry response.

But here’s the problem. We don’t lack a coalition of the willing, we lack coordination of the willing. Organised crime is getting more organised, but we’re all putting out spot fires in the middle of a bushfire, despite the goodwill among industry players.

For example, Finsure’s proprietary compliance and risk-monitoring tool, Sentinel, leverages AI to identify any suspicious activity, hidden risks or undisclosed relationships between borrowers and referral partners. But our ability to catch risks is limited by lack of access to data held by lenders, aggregators and third parties.

Data is the key.

We run every check available to us at onboarding, but those checks can’t reveal hidden associations or coordinated behaviour across institutions. The data that would expose it sits siloed with lenders, aggregators and third parties. As a result, we’re expected to stop what we can’t see.

We are trying to fight fraud with one hand tied behind our back.

Of course, there are additional steps we can take to mitigate individual cases. For instance, we have partnered with document fraud detection firm Fortiro to combat AI document manipulation through an integration into our upcoming Metanoia CRM platform. They are an organisation that truly understands the challenge we’re facing. In fact, Fortiro CEO Sean Quagliani recently said about 7 per cent of customer loan documents it assessed exhibited “high risk” sign s – this is a significant issue which must be solved.

What is the solution?

There are three things that would really move the needle.

Better data. Real-time bank-to-aggregator data sharing and reference checking, to ensure that any issues or undisclosed relationships between borrowers and referral partners are picked up in real time by AI monitoring tools. This can be done now. To quote William Gibson, “the future is already here, it’s just not evenly distributed yet.”
Call out the real bad guys. A central, governed register of proven bad actors, rather than a secret list of people who might be dodgy. This stops the real bad apples from bobbing up elsewhere in the barrel.
Finally, an alignment on regulation. The recent AUSTRAC investigation, whilst limited in its findings, has put Aggregators and brokers on notice that we may need to align to the anti-money laundering reforms, as we combat this threat against the industry.

I have a serious concern that recent behaviours by a small number of individuals within the financial system have put mortgage brokers on the radar of AUSTRAC and, unless as an industry we work together and towards a common north star, I foresee mortgage brokers becoming a reporting entity under the AML/CTF Act.

If we can make real, meaningful change, the prize is a system that is both harder to defraud and fairer to honest brokers. Secure shared data means shared security.

I look forward to a connected, efficient, data-led industry where good brokers are protected and fraud has nowhere to hide.