The AI Scam Epidemic: How Australians Are Losing Billions — and Why AI Is Now Fighting Back
Australians lose billions of dollars to scams every year, and AI-powered tools — voice cloning, deepfakes, and hyper-personalised phishing — are making the problem worse. But the same technology is now being turned against the scammers.
Australia has a scam problem. The ACCC's Scamwatch reported that Australians lost more than $2.18 billion to scams in 2025, and the methods being used are becoming more sophisticated by the month. But the same artificial intelligence technology that is enabling a new generation of scams is also being deployed in reverse, by banks, telecommunications companies, and government agencies working to detect and stop fraud before it reaches its victims.
This is the story of an arms race — and why, for the first time in years, there are genuine reasons to be cautiously optimistic about the outcome.
The Scale of the Problem
The ACCC publishes an annual Targeting Scams report that draws on data from Scamwatch, the Australian Financial Crimes Exchange, and other government agencies. The 2025 report recorded losses of $2.18 billion across all scam types reported to those agencies. The ACCC has consistently noted that this figure represents only a fraction of actual losses, because the majority of scam victims do not report what happened to them.
The types of scams causing the most harm have shifted significantly over the past three years. Investment scams remain the single largest category by dollar value. AI is increasingly being used to make scams more convincing, personalised and difficult to detect, including through synthetic voices, deepfake videos and AI-generated messages.
Voice cloning — the use of AI to replicate a person's voice from a short audio sample — has made a particularly troubling category of scam more convincing. Scammers have used cloned voices to impersonate family members in distress, calling elderly Australians and asking for urgent financial help. The technology required to do this has become accessible and inexpensive, and the results can be difficult to distinguish from a genuine call.
Deepfake video has followed a similar trajectory. The ACCC's Scamwatch has documented cases where scammers used AI-generated video of public figures — including well-known Australian business identities — to promote fraudulent investment schemes. These videos are distributed through social media platforms and can reach large audiences before they are identified and removed.
How AI Is Being Used Against Scammers
The encouraging development is that the institutions best placed to interrupt scams — banks and telecommunications companies — have been investing heavily in detection and disruption systems, and those systems are beginning to produce measurable results.
Telstra's Cleaner Pipes Programme
Telstra has publicly described its Cleaner Pipes initiative as one of the largest anti-scam programmes operated by any telecommunications company in the world. The programme analyses call and SMS traffic across Telstra's network, identifying patterns associated with scam activity and blocking suspicious communications before they reach customers.
Telstra has published figures on the programme's performance. On average, the programme blocks around 21 million scam calls and 11 million scam SMS messages per month on its network. The company has also deployed tools to identify and block websites associated with phishing and malware, and has shared intelligence with other carriers and government agencies through industry coordination mechanisms.
The Cleaner Pipes programme operates at the network level, which means it can intercept scam communications regardless of which device or application a customer is using. This is a significant advantage over consumer-level protections, which depend on individuals recognising and avoiding suspicious contacts.
NAB's Scam Detection Systems
National Australia Bank has been publicly vocal about its investment in fraud and scam detection. NAB has described systems that analyse transaction data in real time, looking for patterns that suggest a customer may be in the process of being scammed. The bank has also implemented measures that introduce friction into high-risk transactions — requiring customers to confirm additional details or speak with a banker before a payment is processed. NAB has stated publicly that these interventions have prevented significant losses, though it has not published a specific dollar figure for prevented losses in a form that can be independently verified.
NAB has also been a prominent participant in the Scam-Safe Accord, a voluntary industry commitment signed by Australian banks in November 2023. The Accord, coordinated by the Australian Banking Association, commits signatory banks to a set of anti-scam measures including confirmation of payee systems, limits on high-risk payment channels, and investment in customer education.
The Australian Government's National Anti-Scam Centre
The Australian Government established the National Anti-Scam Centre (NASC) within the ACCC in July 2023. The NASC coordinates intelligence sharing between government agencies, financial institutions, and telecommunications companies, and operates a disruption programme that targets the infrastructure used by scammers — including websites, phone numbers, and bank accounts.
The NASC has developed a "fusion cell" model: time-limited taskforces that bring together government, banks, telecommunications companies, digital platforms and other organisations to tackle specific scam types. These taskforces allow participants to share intelligence and coordinate disruption efforts more quickly than traditional siloed approaches permitted.
The AI Arms Race: Why This Is Not a Simple Victory
It would be misleading to suggest that AI is simply solving the scam problem. Detection systems must continually adapt as scammers change their tactics, platforms and payment methods. When scammers shift to new channels or new social engineering scripts, detection systems need time to catch up. This creates windows of vulnerability that sophisticated scam operations can exploit.
There is also a significant challenge around the international nature of scam operations. Many of the criminal organisations running large-scale scam campaigns operate from jurisdictions where Australian law enforcement has limited reach. The disruption of scam infrastructure — taking down websites, blocking phone numbers, freezing accounts — is effective but not permanent; scammers can rebuild their infrastructure relatively quickly.
What Australians Can Do
While the institutional response to scams is improving, individual awareness remains an important layer of protection. The ACCC's Scamwatch publishes current information about active scam types and provides a reporting mechanism that feeds directly into the NASC's intelligence systems.
Several practical steps are consistently recommended by Scamwatch and the major banks:
Verify unexpected contact independently. If you receive a call, message, or email from someone claiming to be your bank, a government agency, or a family member in distress, hang up and call back on a number you find independently — from the organisation's official website or a number you already have saved. Do not use contact details provided in the suspicious communication.
Be sceptical of investment opportunities promoted through social media. The ACCC has specifically warned about AI-generated video advertisements featuring well-known Australians promoting investment schemes. If an investment opportunity appears in a social media advertisement, treat it with significant scepticism regardless of who appears to be endorsing it.
Use your bank's Confirmation of Payee tools. Confirmation of Payee began rolling out nationally in July 2025 and is now available at participating banks. These systems verify whether the name on a bank account matches the account number you are paying, providing a meaningful check against misdirected payments.
Report scams to Scamwatch. Every report contributes to the intelligence picture that the NASC uses to identify and disrupt scam operations. Reporting takes a few minutes and directly supports the national response.
A More Sophisticated Response
Australia's anti-scam response is becoming substantially more sophisticated, although reported losses remain above $2 billion a year and scammers continue to adapt. The investment in detection and disruption infrastructure — from Telstra's network-level blocking, to NAB's transaction monitoring, to the establishment of the NASC — represents a qualitative shift in Australia's capacity to respond.
The Scam-Safe Accord commits the banking sector to a set of measures that, when fully implemented, will make it harder for scam proceeds to move through the Australian financial system. The NASC's fusion cell model is building the kind of coordinated, cross-sector response that has historically been absent from the Australian approach.
None of this eliminates the risk. Scammers are adaptive, well-resourced, and operating at scale. But the institutions with the greatest capacity to interrupt scams are deploying tools and coordination mechanisms that better match the sophistication of the threat — and that is a meaningful development for every Australian.
Stay informed
Get AI news every Friday
The AI Digest delivers the week's most important AI stories — free, in plain English.
Subscribe free →Related Articles
More AI in Australia →How AI Is Giving Australian Farmers an Early Warning on Drought
Australian farmers are gaining access to earlier, more actionable drought warnings through AI tools built on CSIRO research and Bureau of Meteorology seasonal forecasting data — changing how they manage stock, water, and forward contracts before dry conditions lock in.
Driverless Cabs Are Coming to Australia: What the Trials, the Technology, and the Road Rules Tell Us
Autonomous vehicle trials are running across multiple Australian states, regulators are actively updating the road rules, and the technology is maturing fast. Here is a clear-eyed look at where driverless cabs stand in Australia right now — and why the outlook is genuinely encouraging.
Australia's AI Skills Gap: Why We're Training for Jobs That Already Exist Instead of the Ones Coming Next
Australia's education system is producing AI-literate graduates — but the skills being taught are already a step behind what employers actually need. Here is what the labour market data is signalling, and what a forward-looking national skills strategy would look like.