Deepfake Albanese Pitching $40,000 Returns Led $7.4M in Australian Scam Losses

Australian Prime Minister Anthony Albanese attends a joint press conference with Vietnamese General Secretary and President To Lam (not pictured) at Parliament House in Canberra on August 11, 2026. HILARY WARDHAUGH/AFP via Getty Images

Australia’s corporate regulator revealed Sunday that victims of scams impersonating the country’s ten most recognizable public figures — including Prime Minister Anthony Albanese — have collectively lost AUD $7.4 million (approximately $5.3 million USD) to deepfake investment fraud, even as the Australian Securities and Investments Commission removed a record 19,400 fraudulent websites and online scams in its most recent fiscal year, nearly three times the volume of the prior period.

Albanese Promised $40,000 Returns. He Never Said It.

The most documented example from ASIC’s latest enforcement update features authentic news footage of Prime Minister Albanese, re-voiced with AI-synthesized audio. In the fabricated clip, a version of the Prime Minister that looks entirely real tells viewers they can turn an AUD $4,000 (approximately $2,848 USD) investment into AUD $40,000 (approximately $28,480 USD) a month through a platform he describes as “official” and “guaranteed by the government.” No such platform exists and Albanese made no such statement.

This construction — real video preserved intact, only the audio track replaced — is technically distinct from the full synthetic deepfakes that get the most public attention, and it is, for that reason, more dangerous than the kind consumer protection guides describe. Standard advice tells viewers to watch for unnatural blinking, mismatched lip-sync, or inconsistent shadows. Those artifacts only appear when a face is artificially generated. In the Albanese scam, the face is real. The facial movements are real. The context — a genuine press appearance, a real setting — is real. Only the audio reveals the fraud, and modern voice cloning tools can generate a convincing replica from as little as three seconds of reference audio — freely available from any public figure’s press conference, interview, or media appearance.

ASIC Chair Sarah Court identified this shift directly. “These scams are targeting all of us, and they are harder and harder to detect,” she told SBS News on Sunday. “Even people that think that they’re reasonably financially savvy — we are seeing examples of people that have been persuaded or lured in by what are very clever attempts by scammers to get money.”

ASIC Removed 19,400 Fraud Sites — and the Scams Kept Coming

The enforcement data ASIC released Sunday illustrates both the scale of the effort and its structural limitation. In its most recent fiscal year, the regulator removed 5,476 phishing scam hyperlinks — an increase of roughly 279% — and 7,051 fake investment platforms, more than 150% more than the prior period. Cryptocurrency scams accounted for 3,106 of the takedowns, a roughly 30% year-on-year rise. Since launching its scam disruption service in 2023, ASIC has removed more than 33,400 malicious investment and phishing sites in total.

At its operational peak, the regulator was removing an average of 32 fraudulent sites per day. The enforcement effort has recently expanded to cover social media advertising — a significant development, because the primary vector for victim acquisition in current scam operations is not a standalone website but a paid ad in a social media feed.

The regulator also documented a technique called “cloaking,” in which scammers display compliant, harmless content to platform content moderators and regulators while serving fraudulent investment ads to targeted users, based on their device or geographic location. The implication is structural: platform content review, even when conducted in good faith, cannot reliably detect ads that detect the reviewer and switch faces.

Australian investment scam losses appear to be declining in aggregate — from AUD $291 million (approximately $207 million USD) in 2023 to AUD $192 million (approximately $137 million USD) in 2024, with AUD $73 million (approximately $52 million USD) reported so far in 2026. Whether that trend reflects enforcement and public awareness, a shift in scammer targeting, or simply underreporting is not yet established. What is clear is that the technology available to scammers is growing more sophisticated at the same time.

“This Has Dramatically Destroyed My Life”

The AUD $7.4 million (approximately $5.3 million USD) figure from ASIC’s update traces to real, named people. According to ASIC’s August 17 update, Kristy Holden, 56, lost AUD $150,000 (approximately $107,000 USD) to a scam exploiting the identity of Commonwealth Bank investment commentator Tom Piotrowski. She is now living in her granddaughter’s bedroom. According to ASIC’s enforcement report, Toni, a 74-year-old retiree who declined to give her surname, was drawn in by a Facebook advertisement in September 2025 and lost AUD $320,000 (approximately $228,000 USD).

“This has dramatically destroyed my life as I am 74, cannot work and now struggle to live on a pension,” Toni said.

The operational playbook these scams follow has been documented by ASIC and the NZ Financial Markets Authority. A victim encounters an AI-generated advertisement or fabricated news article on social media, typically showing a trusted figure promising extraordinary, low-risk returns. They are directed to a convincing fake investment platform that shows fabricated profits accumulating in their account. When they try to withdraw, they are asked to pay additional fees. The money does not return.

The scam extends beyond politicians to economists, financial journalists, and business figures. Among the most commonly cloned identities ASIC identified were economists Tom Piotrowski and Stephen Koukoulas, ABC financial journalist Alan Kohler, politicians Jacqui Lambie and Angus Taylor, mining magnate Gina Rinehart, One Nation leader Pauline Hanson, and businessman Dick Smith. In one documented scam, a fabricated Nine News broadcast segment showed presenters Deb Knight and Chris Kohler endorsing a trading platform. In another, a deepfake of Koukoulas circulated in WhatsApp communities promoting a stock tip.

Why Trusted Faces Work — and What That Means for Consumer Defense

Psychologist Robert Cialdini identified social proof psychology in his foundational 1984 work, “Influence: The Psychology of Persuasion,” as one of the six primary principles of human persuasion: individuals in uncertain situations defer to the behavior or endorsement of people they respect. In investment decisions — inherently uncertain, high-stakes, and involving unfamiliar products — the social proof value of a trusted face is disproportionately high.

Deepfake technology specifically weaponizes this mechanism. A fraudulent platform introduced by an anonymous email requires the victim to overcome distrust. The same platform introduced by what appears to be the sitting Prime Minister, in what appears to be an authentic news setting, provides three reinforcing trust signals simultaneously: the face, the context, and the authoritative vocal register. When that voice is synthesized from a real person’s audio profile, the cognitive signal is genuine. The deception occurs below the level where deliberate evaluation normally operates.

Court’s advice points to the same conclusion from a consumer protection angle: “The presence of polished content, familiar branding or convincing testimonials does not mean an investment is legitimate. AI is making investment scams more convincing and harder to detect. A simple online search is not enough to verify whether an opportunity is legitimate.”

McAfee’s 2025 research found that 72% of Americans had seen fake celebrity or influencer endorsements online, and 39% had clicked on one — with 10% of those who clicked ultimately losing money, averaging $525 per incident. A January 2026 report by blockchain analytics firm Chainalysis found that trading platform impersonation scams grew more than 1,400% year-over-year, with AI-enabled operations extracting 4.5 times more money per victim than traditional fraud methods.

Is This Pattern Specific to Australia?

It is not. In April 2026, New Zealand’s Financial Markets Authority reported identifying 110 scam advertisements published on Meta platforms within a single 24-hour period, while flagging more than 190 fake trading sites for removal since March 2026. The FMA documented a nearly identical operational playbook: deepfake video of a trusted public figure endorsing a fake investment opportunity, linked to a fabricated news article carrying stolen logos from RNZ, TVNZ, or the NZ Herald, directing victims to a fake platform. A Taranaki grandmother lost NZD $224,000 to a fake deepfake of NZ Prime Minister Christopher Luxon promoting Bitcoin.

Globally, cryptocurrency scams involving AI deepfakes cost consumers and businesses an estimated USD $4.6 billion in 2024, with deepfakes accounting for nearly 40% of high-value fraud, according to a joint report by crypto exchange Bitget and security firms SlowMist and Elliptic. The US Commodity Futures Trading Commission warned in late 2025 that deepfake video and voice cloning were being used in live video calls to impersonate licensed brokers in real time.

What ASIC Wants Australians to Do

ASIC’s recommended consumer protections are specific and actionable:

Verify the Australian Financial Services (AFS) license number independently through ASIC’s free public registers before committing any money. If a business cannot produce verifiable license details, or if those details do not match ASIC’s records, do not invest. The verification tool is available at moneysmart.gov.au.

Check website addresses carefully. Scam sites routinely register domains that differ from legitimate ones by a single character. Be wary of urgency — phrases like “one-time offer,” “invest now,” and “limited availability” are operational signals in the scam playbook, not legitimate sales tactics. Do not trust testimonials or endorsements, however polished. Do not assume a social media advertisement has been vetted by the platform on which it appears.

Court summed up the verification requirement plainly: “Before investing, consumers should verify website addresses, check whether a person or company is legitimate and who they claim to be, and be wary of urgent calls to act. If a business cannot provide verifiable license details, or if the information does not match ASIC’s records, do not invest.”

Suspected scams can be reported to Scamwatch at scamwatch.gov.au.

Why Enforcement Alone Has Limits

ASIC’s removal effort is both the largest such program in Australia’s regulatory history and, structurally, a defensive measure against a self-renewing problem. Each website removed can be replaced by a new one — sometimes within hours — by operators who have automated the process of cloning fraudulent site templates.

The contrast with Taiwan is instructive. Taiwan’s government mandated that social media platforms verify the identity of financial-product advertisers before their ads run. Meta, which had initially sought regulatory extensions, ultimately complied under the threat of commercially significant penalties. The result, per Taiwan’s Ministry of Digital Affairs: investment scam ads fell 99%, and identity impersonation ads dropped 98%. Meta’s own internal analysis had anticipated that universal verification would meaningfully reduce scam ad prevalence before the Taiwan mandate took effect. Australia does not currently have equivalent legislation requiring platforms to verify the identity of financial advertisers before ad placement. ASIC’s enforcement model operates after the scam ad is running — after a victim may already have seen it.

For individual Australians, ASIC’s update carries an uncomfortable practical conclusion: in the era of AI voice synthesis, a face you recognize in a video and a voice that sounds exactly like that person’s are no longer reliable indicators that the person is actually speaking. The moneysmart.gov.au license check is the verification tool that synthetic media cannot replicate.

Frequently Asked Questions

How do deepfake investment scams targeting Australian public figures actually work?

The scam documented by ASIC in its August 17, 2026 update uses a technique called audio grafting: authentic video footage of a recognizable public figure — in the most prominent example, Prime Minister Albanese — is preserved intact while only the audio track is replaced with AI-synthesized speech. Researchers suggest that a voice clone requires as little as three seconds of reference audio, and that material is freely available for any public figure who has given interviews, press conferences, or media appearances. The resulting video looks entirely authentic because the face and facial movements are real. Only the voice has been replaced. Standard detection advice — look for unnatural blinking or mismatched lip-sync — does not apply to audio-grafted deepfakes.

If I see an investment ad featuring a famous Australian on social media, how do I check if it is real?

ASIC’s direct guidance: check whether the person or company in the ad holds a current Australian Financial Services (AFS) license through ASIC’s free public register at moneysmart.gov.au. This is the one verification step that a deepfake cannot replicate, because a fictitious investment platform will either have no AFS license or will misrepresent an existing one. If the license details do not match ASIC’s records, do not invest and do not provide personal information.

Are investment scam losses in Australia actually getting better or worse overall?

The aggregate trend shows declining reported losses — from AUD $291 million (approximately $207 million USD) in 2023 to AUD $192 million (approximately $137 million USD) in 2024, with AUD $73 million (approximately $52 million USD) reported so far in 2026 — which suggests enforcement and public awareness are having some effect. However, ASIC’s data also shows the sophistication and volume of attacks increasing: 19,400 sites removed in the most recent fiscal year, nearly three times the prior-year figure, and AI audio-grafting techniques now producing scams that pass standard visual detection. A declining loss trend alongside rising attack sophistication is not a settled resolution.

Are deepfake investment scams using politicians unique to Australia?

No. New Zealand’s Financial Markets Authority documented 110 deepfake investment scam ads on Meta in a single day in April 2026, impersonating NZ public figures including Prime Minister Christopher Luxon and deputy prime minister Winston Peters. Japan saw ¥127.4 billion (approximately $804 million USD) in social media investment fraud losses in 2025, prompting a joint demand from seven Japanese cabinet agencies to major platforms. The United States CFTC warned in late 2025 about live video calls using deepfake brokers. The playbook is the same across jurisdictions: a trusted local face, a fake investment platform, and a social media advertising system that in most countries does not require financial advertisers to verify their identity before purchasing ad placement.