Scam losses hit AUD $158.5M as Arctic Wolf warns of AI fraud ahead of Black Friday
Arctic Wolf warns Australian retailers of AI-driven fraud before Black Friday as scam losses hit AUD $158.5 million and 750 malicious holiday domains are confirmed.

Why it matters
- Australian scam losses from websites, apps and social media reached AUD $158.5 million in 2025, up 21.8% year over year
- Fortinet FortiGuard Labs found over 18,000 holiday-themed domains registered before the 2025 season; at least 750 confirmed malicious
- Stripe reported a 38-fold monthly increase in AI-agent purchases via its Link platform in September 2026
- Arctic Wolf identified five threats: fake retailers, malware-laced verification requests, account takeovers, AI scams, and AI shopping-agent risks
- David Jones entered voluntary administration amid broader pressure on Australian retail
Australian consumers lost AUD $158.5 million to scams initiated through websites, mobile apps and social media in 2025, a 21.8% rise on the previous year — and cyber security firm Arctic Wolf says Black Friday will amplify the threat for retailers and shoppers alike.
The firm this week warned Australian retailers of a surge in cyber scams and AI-linked fraud heading into the peak shopping period. It identified five main threats: fake retailers and brand impersonation, fraudulent verification requests carrying malware, account takeovers, AI-assisted scams, and risks tied to personal AI shopping agents.
The stakes are unusually high this year. Australian retail is under severe strain, and Arctic Wolf pointed to the recent voluntary administration of David Jones — one of the country's best-known department store chains — as evidence of the pressure. For merchants fighting for survival, a fraud incident or a wave of chargebacks during the biggest sales window of the year compounds an already difficult trading backdrop.
How big is the fraud surface?
The numbers around holiday-season fraud are concrete. Fortinet's FortiGuard Labs found that more than 18,000 holiday-themed domains were registered in the three months leading up to the 2025 holiday season. At least 750 of them were confirmed as malicious.
That figure only captures confirmed-bad domains. The gap between 750 confirmed malicious and 18,000 newly registered suggests many more remain unclassified — a caveat worth holding onto when reading any single fraud statistic.
The National Anti-Scam Centre's AUD $158.5 million loss figure covers scams initiated through websites, apps and social media. It does not capture the full cost to merchants: diverted sales, brand damage from impersonation, and refund disputes sit outside it.
What scams will shoppers actually see?
Arctic Wolf said consumers are most likely to encounter:
- fraudulent retailers and counterfeit products
- deceptive adverts
- phishing messages
- false delivery notifications
These scams aim to collect payments, credentials or personal information while diverting customers away from legitimate retail channels. The National Anti-Scam Centre has separately warned that increasingly sophisticated scams, including those using AI, are becoming harder for consumers to recognise.
A second, more technically aggressive pattern is also circulating. Arctic Wolf flagged the use of fake CAPTCHA tests, browser errors and verification prompts that persuade users to run malicious commands on their own devices. The method is known as ClickFix (a social-engineering technique that tricks users into pasting commands into their own terminal), and Arctic Wolf has investigated campaigns using it. Microsoft has reported similar tactics targeting thousands of enterprise and consumer devices globally every day.
Black Friday itself provides cover for these attacks. A prompt asking a shopper to verify their identity, activate a coupon, fix a browser error or resolve a delivery problem seems plausible during a period of heavy online shopping and shipping activity.
Why are compromised accounts worth more during a sale?
Account security is a central concern in Arctic Wolf's advisory. Compromised email and retail accounts let attackers reset passwords, intercept order confirmations, redirect deliveries, and access personal or financial information.
The firm also warned that stolen browser cookies and authentication tokens (small files and keys that keep a user logged in) can allow attackers to hijack existing sessions — in some cases bypassing password-based controls entirely.
During major sales events, compromised accounts carry extra value. They can include saved payment cards, gift-card balances, loyalty points, delivery addresses and complete order histories. A single account takeover during Black Friday can therefore be more lucrative for a criminal than the same attack in a quiet trading month.
How is AI changing the scam playbook?
Arctic Wolf said AI-generated text, images, adverts and fake profiles are making scams harder to spot. The tools let criminals build localised campaigns that mimic specific retailers, products and delivery providers, stripping out the obvious warning signs — bad spelling, clumsy design — that once exposed fraudulent operations.
The firm advised consumers to treat steep discounts, urgent messages, unfamiliar website addresses, and requests for unusual payment methods or software downloads as red flags.
A newer and less understood risk comes from personal AI shopping agents — automated tools that browse products and make purchases on a user's behalf. Stripe reported a 38-fold monthly increase in purchases made through its Link platform using AI agents in September 2026, though it did not disclose the underlying transaction volume. A 38-fold rise on an undisclosed base tells you the direction of travel, not the scale.
Justin Moore, Director, Adversary Operations at Arctic Wolf, said the dual shift — in consumer behaviour and criminal tactics — has changed the threat landscape. "AI is changing both how consumers shop and how cybercriminals operate. Shoppers can no longer rely on obvious spelling mistakes or poorly designed websites to identify scams. Retailers face the growing challenge of distinguishing genuine customers and authorised AI shopping agents from malicious bots," Moore said.
For shoppers, the agent trend carries a subtle risk: a user may approve a payment without realising the agent is buying from a fraudulent seller or purchasing counterfeit goods. The approval step exists, but the scrutiny behind it may not.
What should retailers do?
Arctic Wolf urged retailers to put tighter checks around AI-driven purchasing activity. Its specific recommendations:
- separate verified AI shopping agents from unidentified automation and malicious bots
- apply stronger authentication checks for gift cards, loyalty redemptions and larger purchases
- monitor for signs of ClickFix-style activity, including browsers launching PowerShell, Command Prompt or other terminal tools
- limit what AI systems can access, treating reviews, emails and uploaded documents as untrusted inputs
- restrict permissions to only what is necessary
Moore said smaller businesses are especially exposed during peak sales periods. "For smaller businesses and brands increasingly reliant on sales during Black Friday, every sale counts. Cybersecurity can quickly fall off the priority list, leaving businesses and customers exposed to scams, fraud and cyber threats," Moore said.
He argued that speed of response now matters as much as prevention. "Black Friday should be an opportunity for retailers to drive sales and for consumers to secure a bargain, not an opportunity for cybercriminals. As scams become more sophisticated and AI changes the way people shop online, businesses need to be able to identify suspicious activity and respond quickly. Protecting customer information and maintaining consumer trust is just as important as making the sale," Moore said.
The warning lands at a moment when the burden of fraud detection is shifting. AI has removed the crude visual cues consumers once used to spot scams, while simultaneously creating a new class of automated buyers that retailers cannot easily classify as friend or foe. How merchants separate authorised AI agents from malicious bots — without adding enough friction to lose the sale — will define the security problem for the next several shopping seasons.
Source: itbrief.com.au
More from James Calloway
Show full bio
News editor covering people, launches and day-to-day practice in artificial intelligence and machine learning at AI In Context.
749 articles
Related articles
- Akamai Warns AI-Powered Scams Target $7.2 Billion Black Friday Shopping Rush
- Australia Warns AI Will Industrialise Scams as Losses Hit $2.18 Billion
- Anomali warns AI agents will run entire fraud campaigns against Australian firms in 2027
- Adactin maps six fraud defences for Australian businesses facing AI scams
- Meta Claims Millions Shop With Muse. Executives See 'Hundreds' of Orders