Subscribe
How AI mortgage fraud could hit honest borrowers in the hip pocket
Costs involved in securing a loan could rise as lenders battle against fraudulent applications.
- Securing a loan could be more expensive as banks are forced to defend against the growing use of AI
- Bank statements topped the list of documents being altered, followed by employment records, pay slips, and proof-of-income paperwork.
- Brokers warn that consumers will ultimately pay if higher costs become the norm.
Borrowers are being warned that securing a loan could get more expensive and require even more evidence as lenders are forced to weed out AI-generated false documents.
Westpac has admitted that fraud controls may need to change, while Commonwealth Bank spent $900 million on fraud and cyber protection. NAB and ANZ have also admitted they are investing heavily because of the heightened fraud environment. Â
How lenders will apply greater costs is not yet known. Loan application fees usually cost between $150 and $990, averaging $468, according to Canstar.
Younger Australians are far more likely to use AI to generate false documentation for mortgage loan applications, exposing how problematic the technology can be for lenders.
The warning comes as national polling finds one in five Australian borrowers have used AI to create or alter documents for a home loan application.
The national poll commissioned by comparison platform OurTop10 found that bank statements topped the list of documents being altered, followed by employment records, pay slips, and proof-of-income paperwork.
AI fraud
AI has shifted mortgage fraud from isolated incidents to an industrialised business model favoured by organised crime syndicates. This will force banks to invest in stronger verification and fraud-detection technology, accounting and consulting firm EY warned in August.
While banks have been scrambling to get a handle on this new risk profile, media reports suggest lenders have been reviewing broker relationships as part of the push to improve fraud detection measures.
Brokers warn that additional compliance will lead to more loans being registered as defaults and higher compliance and administrative costs, which consumers will eventually pay for.
Commonwealth Bank blew the whistle earlier this year, suspecting around $1 billion in home loans were obtained fraudulently, including through AI-generated documents.
Austrac’s Fintel Alliance also revealed coordinated mortgage fraud and systematic weakness across the nation’s lending sector in September. Â
Austrac CEO Brendan Thomas said the findings were a “wake-up call for lenders” that demonstrated the importance of identifying fraud before approving a loan.
“The most effective way to stop mortgage fraud is before a loan is approved. Once a loan is established and the funds have moved, recovering the money becomes significantly harder,” Thomas said in a statement.
Fraud brings penalties
Borrowers who falsify documents must be aware they are committing fraud, which carries serious penalties, Sydney Mortgage Choice principal and broker Richard Brown told The Money Edition.
“As a broker, providing a defence against financial fraud is becoming a bigger part of our business,” Brown said.
He said he was concerned that even with the checks and double checks brokers undertake, the quality of some AI-doctored documents may be difficult for brokers to spot.
“This means a broker could be unknowingly putting their license and business at risk,” he said.
Brown warned consumers will ultimately pay if higher costs become the norm.
“Falsifying income documents means borrowers will potentially get a loan they cannot afford, which could put them in financial hardship, not to mention that there are serious ramifications with these actions,” he said.
“We may be headed quickly towards a position where lenders don’t accept client-supplied documentation. Instead, reliance on open banking and ATO data sharing may become the verification that lenders require.”
Fellow Sydney mortgage broker Rebecca Jarrett-Dalton told The Money Edition she has been fielding a growing number of enquiries based on a recommendation from ChatGPT, often coming with a list of pre-prepared questions.
She said that brokers need to be cautious to protect sensitive data.Â
“We actually have an AI interface available that intuitively collects clients’ details, but very few are using it now — and essentially that’s the only way a client can use AI for this process where they are involving a lender,” she said.
Meanwhile, the big four banks are requesting access to Taxation Office data to verify income details.
While the government would need to amend the Taxation Administration Act to make that happen, this year’s Federal Budget allocated millions to build infrastructure to investigate open banking reforms, which would enable data sharing.
The Australian Banking Association is pushing for better data sharing, arguing that the lack of data sharing in the digital age creates unnecessary risk in the financial system.