Fresh figures have revealed criminals are harnessing artificial intelligence to embed billions of dollars of bogus loans in bank portfolios worldwide.
As Australian authorities confront an alleged wave of mortgage fraud worth at least $1 billion, linked to doctored income documents, new research has revealed that lenders across the globe are fighting similar threats.
The latest Global Credit Trends 2025 report from credit reporting agency Equifax has detailed how AI‑enabled fraud rings are using synthetic identities and digital loopholes to hide more than a billion dollars of bad debt in Canada alone.
AI pushes Canadian credit fraud into ‘hidden’ territory
Equifax’s analysis of Canadian consumer credit data found that by the end of 2025, lenders were carrying about $1.3 billion in suspected “hidden fraud” within 90‑plus‑day past‑due balances across credit cards, unsecured lines, and auto loans.
This is debt that presents in company systems as ordinary arrears but is believed to stem from deliberate fraud, rather than genuine borrowers in hardship.
In its report, the credit bureau warned that organised crime had upgraded its methods, using advanced technology to probe weak spots in lenders’ controls.
“Credit fraud in Canada is becoming increasingly sophisticated as organised crime groups leverage digital gaps and AI to exploit the financial system,” Equifax said.
“In 2025, we saw a significant rise in vehicle-related fraud and ‘faceless’ identity crimes that may bypass traditional security.”
It said auto finance had become a particular flashpoint in Canada, with the most common schemes now involving fraudsters exploiting reporting lags and fragmented data flows between lenders.
“The most prevalent forms of credit fraud is auto loan stacking, where criminals exploit reporting delays to finance multiple vehicles simultaneously with no intent to repay,” the report read.
“Never pay fraud, involving the opening of new accounts followed by immediate, total default; and ‘bust out’ fraud, a ‘long con’ where fraudsters build strong profiles only to max out all limits and vanish.”
Underpinning much of this alleged activity is synthetic identity fraud, in which real personal data is fused with fabricated details to create borrowers who look legitimate on paper.
Equifax stressed that this technique allowed fraudsters to operate for long periods without attracting the attention of the actual victims whose information had been partially stolen.
“Synthetic identity fraud often acts as the engine for these schemes – blending real stolen data with fake details to execute these high-value operations without a real victim ever being alerted,” Equifax said.
The bureau’s estimates suggest that “hidden fraud” embedded in 90‑day‑plus balances at the end of 2025 totalled $55 million in credit cards, $141 million in unsecured lines, $570 million in auto loans, and $136 million in telco debt.
Those balances, while already seriously delinquent, may still be treated by lenders as recoverable customer exposures rather than outright fraud losses.
Equifax further warned that this misclassification had implications well beyond individual write‑offs.
“Lenders are increasingly exposed to financial losses from ‘hidden’ fraud within their portfolios, which can go unidentified and lead to ineffective, reactive adjudication strategy decisions, such as tightening credit score cut-offs that ultimately hamper portfolio growth,” it said.
It said that kind of blunt tightening could also distort pricing, reduce access to credit for genuine higher‑risk borrowers, and complicate risk‑based capital planning.
Operational impacts are also mounting as collections teams chase debts that will likely never be repaid.
“Collections and recovery strategies may also be impeded, resulting in wasted operational activity and resources spent attempting to recover outstanding balances from fraudulent accounts,” Equifax said.
Equifax said that, rather than relying solely on manual reviews, lenders needed to adopt new tools that could separate synthetic and first‑party fraud from legitimate delinquency early in the credit life cycle.
The firm urged creditors to deploy more sophisticated analytics and machine‑learning tools to fight back against AI‑enabled fraudsters.
“Leveraging targeted fraud tackling scores and AI models combined with improved collections and recovery segmentation can assist in lenders growing their portfolios while mitigating this risk,” Equifax said.
Australian lenders and regulators wrestle with escalating mortgage fraud
The issue comes as banks across the globe, including Australia, uncover billions of dollars of potential loan fraud.
On 27 February, mainstream media reported that the Commonwealth of Bank of Australia (CBA) had self-reported to the police and corporate regulators regarding potential mortgage fraud totalling an estimated $1 billion.
The Australian Securities & Investments Commission (ASIC) later told a parliamentary joint committee on 9 March that it was currently undertaking “compliance inquiries” with Australia’s largest lender in relation to the matter.
The concerns appear to be focused on money laundering – whereby criminals may be ‘washing’ money obtained through illegal means (for example, drug dealing, prostitution rackets, or modern slavery) by obtaining loans using doctored documents – in some cases, created by artificial intelligence (AI).
According to The Australian Financial Review reports, the loans allegedly included “fake income statements” provided by “several accountants” to brokers and were written through both the broker channel and the bank’s controversial introducer program.
It was then revealed on Friday (10 April) that aggregation group Finsure had been implicated in allegations relating to suspected mortgage fraud, with bankers suggesting ‘people within its network’ may be involved in fraudulent home loan applications.
In a statement to The Adviser, Finsure CEO Simon Bednar noted that it had “not been directly contacted by any lenders or regulators regarding the current review being undertaken” and had no details on the allegations.
However, the Australian Transaction Reports and Analysis Centre (AUSTRAC) is currently working with major banks, law enforcement, and other regulators to map the extent of increasingly complex loan fraud, with its CEO Brendan Thomas telling The Adviser that the watchdog was assessing how widespread the issue had become.
Mortgage and Finance Association of Australia CEO Anja Pannek told The Adviser on Monday (13 April) that mortgage fraud was a serious matter – but was not representative of the vast majority of participants in the lending ecosystem.
“It’s not a bank, broker or referrer only issue, it’s a whole-of-industry challenge that requires a co-ordinated response,” she told The Adviser.
Finance industry associations, including the MFAA, wrote to the federal Treasurer Jim Chalmers calling on the government to expand the Consumer Data Right (CDR) to include secure, consent-based access to ATO-held notices of assessment (NOAs) and income tax return information, as well as ASIC company registry data to reduce the risk of mortgage fraud.
[Related: Major aggregator flagged in mortgage fraud probe]
Want to see more stories from trusted news sources?
Make The Adviser a preferred news source on Google.
Click here to add The Adviser as a preferred news source.
