Fake Faces. Real Money.

 

INVESTIGATION  /  TECH & CRIME

Fake Faces. Real Money.

Who Is Profiting From India’s Deepfake Scam Economy?

A fake celebrity endorsement may be the bait. The real business is a chain of ads, private chat groups, fabricated profits and mule bank accounts. India’s cyber-fraud numbers show the scale of the problem—but not yet how much of it is powered by deepfakes.

By Murthy ASR  |  Biz Masala  |  13 October 2026

Editorial illustration: a synthetic face can open the door; the money trail runs through platforms and financial accounts.

The lie begins with a familiar face

The video looks convincing. A public figure appears to recommend an investment platform. The voice, expression and confident delivery seem familiar. A viewer clicks the link, joins a private group and is invited to start small. A dashboard shows a gain. The next deposit is larger. When the investor tries to withdraw, a new demand appears: tax, verification, processing or an “unlocking” fee.

That sequence is central to cases reported by Hyderabad police: AI-generated or manipulated endorsements can lend borrowed credibility to fraudulent investment pitches. In July 2026, Hyderabad cybercrime police summoned Meta and sought details of safeguards across Facebook, WhatsApp and Instagram after complaints involving fake videos of public figures promoting dubious investment platforms. The company’s controls and response should be assessed from records and its own statements, not assumed.

The fraud does not depend on a perfect deepfake. It depends on getting a potential victim to trust a message long enough to transfer money—and then keeping that person engaged until the money is difficult to recover.

The number behind the noise

Government figures published by the Press Information Bureau put reported cyber-fraud losses at ₹22,495 crore in 2025, compared with ₹22,848 crore in 2024. The 2025 total came from about 24.03 lakh financial-fraud complaints. In 2021, reported losses were ₹551 crore. These figures cover financial cyber fraud broadly; they are not a measure of deepfake fraud alone.

Source: Ministry of Home Affairs figures reproduced by PIB. Values are reported financial cyber-fraud losses, not deepfake-only losses.

The total eased slightly in 2025 but remained more than forty times the 2021 figure.

A separate government presentation cited by PIB in August 2026 said fake investment and trading scams accounted for 76% of reported financial losses in 2025. That category includes Ponzi schemes, fake stock-trading offers and crypto scams; it should not be read as a deepfake statistic.

Hyderabad: a local view of a national problem

Hyderabad residents lost ₹17.34 crore to cyber fraud in September 2026 across 464 cases, according to The New Indian Express on 11 October. Police said ₹88.48 lakh was refunded during the month. They also identified 140 Facebook and Instagram profiles linked to 30 paid advertisements promoting a range of misleading schemes. Those ads were not all necessarily deepfake investment scams; the monthly figures cover several kinds of online fraud.

Telangana Cyber Security Bureau director Shikha Goel has said nearly 60% of money lost by cybercrime victims in the state is tied to stock-market and investment-related fraud. In a separate two-week multi-state operation announced in October, the bureau said investigators traced about ₹17 crore in fraudulent transactions and arrested 101 people, including alleged agents and mule-account holders.

These are important enforcement indicators, but they do not tell us how many cases began with a synthetic video, how many advertisements were paid for by the same networks, or how much of the money ultimately reached the masterminds.

How the scam supply chain works

Illustrative pattern synthesised from reported investigation accounts; not every case follows every step.

Biz Masala graphic: the path from paid or viral ad to money movement. The exact route varies by case.

The fake endorsement is the front end. The back end may involve cloned websites, private messaging groups, fabricated trading dashboards, accounts held by intermediaries and rapid transfers across multiple banks or jurisdictions. In a reported Telangana operation, investigators said mule accounts and agents were used to receive, move and withdraw allegedly fraudulent funds. Individual culpability must be established by investigators and courts.

The questions therefore go beyond “Can we detect a fake face?” Who paid to distribute the ad? Who operated the destination website? Which accounts received the money? How quickly were transfers flagged? Where did the proceeds go?

Frozen is not the same as returned

Government data recorded 65,89,201 financial-fraud complaints between 2021 and 2025, with ₹55,050 crore reported lost. A government reply cited by PIB says more than ₹8,189 crore had been saved through the Citizen Financial Cyber Fraud Reporting and Management System by 31 December 2025, across more than 23.61 lakh complaints. Funds held or marked under lien should not be confused with final reimbursement: freezing can be an interim step, and a case may involve several accounts and claimants.

The national helpline 1930 and cybercrime.gov.in are intended to help victims report financial cyber fraud quickly. Immediate reporting can improve the chance of blocking a transfer before funds move again, but it is not a guarantee of recovery.

Platforms, banks and the speed problem

Hyderabad police’s reported summons to Meta raises a practical question: how quickly can a platform detect a fraudulent endorsement, identify the advertiser, remove the content and preserve information for investigators? A video may be removed after a complaint, but copies can be reposted from new accounts or redirect users to encrypted groups.

Banks and payment intermediaries face a different challenge: spotting unusual flows across accounts without treating every legitimate transfer as suspicious. Investigators have described mule accounts as part of the infrastructure used to move proceeds. Account-opening controls, transaction monitoring and rapid inter-bank coordination are therefore central to the response.

Legal deadlines matter, but the original draft’s specific claims about 2026 IT Amendment Rules and three-hour/two-hour deadlines should be checked against the final Gazette notification and the exact scope of each rule before publication. A deadline after a valid order is not the same as proactive detection of every scam ad.

What the public data still cannot answer

·       How much of India’s annual cyber-fraud loss was directly enabled by deepfake video or AI voice cloning?

·       How many scam ads were approved, paid for, rejected, reported and removed—and how quickly?

·       What share of frozen funds was ultimately returned, and how long did restitution take?

·       How many mule accounts were opened through agents, and how many were detected before large transfers?

·       How many complaints led to FIRs, chargesheets, convictions and recovery, using comparable definitions?

Without these answers, the public can see the scale of the losses and some enforcement activity, but cannot fully evaluate the performance of platforms, banks or investigators. Nor can it isolate deepfakes from the wider universe of financial cybercrime.

Five red flags before you transfer a rupee

Reader guide: warning signs to check before transferring money.

·       A famous face promises extraordinary returns. Treat the endorsement as unverified. Check official channels and the firm’s regulator status independently.

·       You are rushed into a WhatsApp or Telegram group. Group members, screenshots and testimonials are not proof that a platform is genuine.

·       A dashboard shows profits you cannot verify. A number on a website is not evidence that assets exist or trades took place.

·       You must pay to release your money. A new “tax”, “unlocking” or “verification” payment is a major warning sign.

·       The beneficiary or account keeps changing. Stop, preserve messages and payment details, and contact your bank immediately.

If you suspect financial cyber fraud, call 1930 as quickly as possible and report at cybercrime.gov.in. Preserve transaction IDs, phone numbers, URLs, screenshots and chat records.

The Biz Masala verdict

A deepfake can counterfeit a face. A scam network converts that counterfeit into trust, and trust into transfers. But the strongest evidence of an industrialised fraud economy is not a convincing video: it is the chain behind it—advertising, recruitment, accounts, transfers and the people coordinating them.

India’s reported cyber-fraud losses are large enough to demand sharper accountability. Yet the national numbers do not isolate deepfake-enabled fraud, and reported arrests or removed ads do not by themselves establish how much damage was prevented. The next step is a measurable public record: ad response times, funds blocked and returned, mule-account disruption, case outcomes and a credible estimate of AI-enabled fraud.

Until that record exists, the safest rule is simple: never transfer money because a video looks familiar. Verify the platform and the person independently.

Sources and reporting notes

·       PIB, 21 Aug 2026 — ₹22,495 crore cyber-fraud losses in 2025; 76% of reported losses attributed to fake investment and trading scams. https://www.pib.gov.in/PressReleasePage.aspx?PRID=2301973&lang=2&reg=48

·       PIB — Government table of financial cyber-fraud complaints and amounts reported, 2021–2025; CFCFRMS funds saved through 31 Dec 2025. https://www.pib.gov.in/PressReleasePage.aspx?PRID=2226441&lang=2&reg=48

·       The New Indian Express, 11 Oct 2026 — Hyderabad September cyber-fraud figures. https://www.newindianexpress.com/cities/hyderabad/2026/Oct/11/hyderabad-residents-lose-over-rs-17-crore-to-cyber-frauds-in-september

·       The Times of India — Hyderabad police summon Meta over AI fake investment scam ads. https://timesofindia.indiatimes.com/city/hyderabad/police-summon-meta-over-ai-fake-investment-scam-ads/amp_articleshow/132719914.cms

·       Hindustan Times, Oct 2026 — Telangana Cyber Security Bureau multi-state operation. https://www.hindustantimes.com/india-news/telangana-cyber-security-bureau-arrests-101-in-multi-state-cybercrime-crackdown-101791368920071.html

·       Telangana Today — Telangana cybercrime losses linked to investment fraud and recovery figures. https://telanganatoday.com/nearly-60-per-cent-of-cybercrime-losses-in-telangana-linked-to-investment-frauds-shikha-goel

·       National Cyber Crime Reporting Portal. https://cybercrime.gov.in

Editorial notes: National cyber-fraud totals cover multiple fraud types and are not deepfake-only figures. Hyderabad September totals also cover several types of cyber fraud. The five-step graphic is an explanatory synthesis, not a claim that every case follows an identical path. Allegations remain allegations unless established by a court. Legal deadlines under the 2026 IT rules require confirmation against the final Gazette text before specific deadlines are stated.

 

Biz Masala

Murthy ASR

Business. Markets. People. Insight.

Illustrations are editorial visuals. They do not depict a verified individual or establish that a specific person committed a crime.