The AI-Fuelled Fraud that Deceived Australia's Biggest Bank

Why It Matters

The case exposes how AI‑enabled document fraud can undermine large‑scale lending, forcing banks and regulators to overhaul verification processes to protect financial stability.

Key Takeaways

• CBA reported over $1 billion in suspected mortgage fraud.
• Fraud leveraged AI‑generated pay slips, shell companies, fake tax returns.
• Referral program abuses allowed brokers and agents to funnel fraudulent loans.
• Investigation expanded to other major banks, potentially another $1 billion.
• Regulators now demand detailed loan data to assess systemic risk.

Summary

The Australian Financial Review uncovered a massive mortgage fraud scheme at Commonwealth Bank (CBA), the nation’s largest home‑lender, that may involve more than $1 billion in illegitimate loans. The scandal emerged after reporters noticed unusual staff turnover in CBA’s home‑loan unit and followed a tip‑box lead that revealed compliance concerns in the bank’s referral program, where accountants, solicitors and real‑estate agents receive commissions for introducing borrowers.

Investigations showed fraudsters used AI‑generated pay slips, fabricated tax returns and shell companies to secure loans, exploiting the referral channel’s weaker due‑diligence controls. Whistle‑blower complaints filed through CBA’s internal SpeakUp system were verified, and the bank ultimately reported the matter to police and the Australian Securities and Investments Commission. The story also linked the fraud to the earlier “Penthouse syndicate” case, prompting CBA to tighten its referral rules and launch an internal review with credit‑bureau Equifax.

"If you ask GPT to create a pay slip it now gives you a ready‑to‑download PDF," said Shaq Quaglani, CEO of fraud‑detection firm Fatiro, highlighting how generative AI accelerates document forgery. Sources confirmed the fraudulent loan book could be as high as $1 billion at CBA alone, with parallel investigations suggesting a comparable exposure across Westpac, NAB, ANZ and other major lenders.

The fallout signals a systemic vulnerability in Australia’s mortgage‑lending ecosystem, prompting regulators such as OSRA to summon ten banks for detailed loan‑book data. Banks face pressure to replace document‑centric verification with data‑driven checks, while the potential seizure of properties financed with illicit funds could reshape credit‑risk management across the sector.

Original Description

Angira Bharadwaj and Joyce Moullakis on why CBA ended up in the middle of a billion dollar loan fraud scandal, how its spreading to the other banks and who might be behind it.

This podcast is sponsored by Woodside Energy (https://www.woodside.com/who-we-are?gad_source=1&gad_campaignid=19236083177&gbraid=0AAAAApAgpoUY9mniozX45G51epawaefJ8&gclid=CjwKCAjwhLPOBhBiEiwA8_wJHCD566vlhh39EEhLaTlb_6HwlTO3T1F04JXFnSdVzfurNzX3B4s9uBoCiMIQAvD_BwE)

Further reading:

Escalating $1b loan fraud scandal threatens to engulf top banks (https://www.afr.com/companies/financial-services/escalating-1b-loan-fraud-scandal-threatens-to-engulf-top-banks-20260311-p5o9im)

AUSTRAC is investigating a widening mortgage loan fraud scandal, after a syndicate duped Commonwealth Bank into writing $1 billion in loans off fake payslips.

AUSTRAC calls in 10 banks for key meeting on spike in mortgage fraud (https://www.afr.com/companies/financial-services/austrac-calls-in-10-banks-for-key-meeting-on-spike-in-mortgage-fraud-20260318-p5okit)

The agency requested data from the lenders as it works to assess the extent of the fraud and whether properties funded by criminal proceeds should be seized.

CBA probes $1b in suspected fraudulent home loans, calls in police (https://www.afr.com/companies/financial-services/cba-probes-1b-in-suspected-fraudulent-home-loans-calls-in-police-20260223-p5o4mc)

The country’s largest bank is working to assess how many loans have been secured based on doctored applications, including documents created using AI.

See omnystudio.com/listener (https://omnystudio.com/listener) for privacy information.