SynthVoice Payday Scam: Disappearing After Fake AI Voice Risk Control and Upfront Fees

Who Gets Targeted

The primary victims are sole proprietors, micro-merchants, e-commerce store owners, and early-stage micro-enterprise owners who generally lack formal financing channels and have a superficial understanding of AI technology while placing high trust in "tech-forward" packaging. These individuals typically face heavy cash flow pressure and short decision-making windows, making them susceptible to pitches such as "no collateral, instant AI approval, same-day disbursement." They often fail to verify platform credentials, gullibly trust polished marketing pages and fabricated success cases, and show insufficient vigilance against upfront-fee scams. Some victims, having previously been rejected by banks, are more inclined to try informal channels, ultimately paying thousands to tens of thousands of yuan in so-called "AI voice model risk control fees" or "voiceprint data modeling fees" before the scammers cut off contact.

骗局怎么运作

• Step 1: Traffic generation via disguised tech loan platforms. Scammers run ads on short-video platforms and WeChat communities, packaging themselves under professional-sounding English names like "SynthVoice Payday" and claiming to provide unsecured, instantly approved loans to micro-enterprises using "self-developed AI voice synthesis and voiceprint risk control models." They forge corporate websites, business licenses, and office scene videos to trick merchants into believing they are legitimate fintech companies.
• Step 2: Inducing merchants to submit basic information and participate in AI voice screening. Customer service guides merchants to fill in their ID cards, business licenses, collection accounts, and other information, then asks them to record a 2 to 5-second voice sample, claiming it is for "AI voiceprint liveness detection and repayment willingness modeling." In reality, this recording may be used to clone voices in subsequent telecom fraud, serving as a tool for secondary crimes.
• Step 3: Collecting upfront fees under the guise of "risk control report fees" and similar names. After the review progress shows "passed AI initial screening," customer service actively contacts the merchant, falsely claiming that due to reasons such as "the AI voice model needing to collect more voiceprint characteristics" or "the need to generate an in-depth risk control credit report," the merchant must first pay a risk control modeling fee or voiceprint data verification service fee ranging from 3,000 to 20,000 yuan, with promises that it will be deducted from the loan principal after disbursement.
• Step 4: Forging AI approval pages and disbursement schedules. After receiving payment, the backend displays "AI final approval passed, funds will arrive within 24 hours," along with a screenshot containing a fake bank transaction number. Scammers stall for time using scripts such as "the AI system is running the last batch of voiceprint comparisons, please be patient," creating the illusion that disbursement is imminent.
• Step 5: Continuing to demand fees or directly cutting off contact under excuses like "data anomalies." When merchants inquire about the disbursement progress, customer service demands additional fees under pretexts such as "AI identified tax data conflicts" or "voiceprint database verification failed requiring a manual review fee." Once the merchant expresses refusal or inability to pay, the platform blocks contact information, dissolves group chats, and shuts down the so-called official website entry, completely disappearing.

红旗信号（看到这些快跑）

• 🚩 Charging fees before loan disbursement: Collecting upfront fees in any form, such as "AI risk control modeling fees," "voiceprint data verification fees," or "credit prepayment processing fees," is a classic scam signal. Legitimate licensed financial institutions do not charge borrowers before disbursing loans.
• 🚩 Demanding voice or video recordings: Requiring loan applicants to record digital readings or specific sentences under the pretext of "voiceprint evaluation" or "AI voice review" may be used to clone voices for secondary fraud and is highly abnormal.
• 🚩 English brand name paired with a crude website: Using international-sounding names like "SynthVoice Payday," but the official website lacks ICP filing information, contact methods are limited to social accounts like WeChat or Telegram, and financial business licenses cannot be queried.
• 🚩 Continuously demanding additional "review fees" or "unlocking fees" before disbursement: When customer service repeatedly demands supplementary fee payments under excuses like "data anomalies" or "AI model stuck orders," payments should be stopped immediately and the police notified.
• 🚩 So-called "instant AI approval" pitches: Legitimate loans require statutory procedures such as credit checks and principal verification. Claims of "no collateral, no credit check, instant AI approval, same-day arrival" are mostly scam pitches and untrustworthy.

真实案例

• In April 2026, a short-video platform exposed user reports against a loan intermediary named "SynthVoice Payday," which collected an 8,000 yuan upfront fee from a small merchant in Zhejiang under the name of "AI voiceprint risk control" and then cut off contact. The victim never received any loans, and the local police station has accepted the case.
• In search results, reports showed that in 2026, multiple regions witnessed AI voiceprint cloning loan fraud cases: strangers lured passersby to read numbers aloud under pretexts like "doing tasks" or "taking photos to test age," and after collecting voiceprints within 5 seconds, impersonated their identities to take out loans on online platforms. Some victims accumulated hundreds of thousands of yuan in debt as a result, prompting police to issue emergency warnings.
• Around June 2026, media outlets such as Jinri Toutiao published early-warning articles titled "Hidden Traps in Silent Stranger Calls: AI Cloned Voiceprints Used for Fraud," disclosing that criminal gangs used AI voice synthesis to impersonate acquaintances or leaders to commit transfer fraud, with victims across multiple regions being defrauded of thousands to tens of thousands of yuan.
• In September 2026, police in Baotou, Inner Mongolia, handled an AI face-swapping and voice-mimicking fraud case. Mr. Guo, a Fujian enterprise head, received a forged video call from an acquaintance who induced him to transfer funds using the company account under the pretext of urgent bidding保证金 of 4.3 million yuan. Mr. Guo transferred 4.3 million yuan in two installments within 10 minutes. Afterward, the police issued an emergency payment stop and recovered a portion of the funds. (Source: [https://new.qq.com/rain/a/20260916A0APD700/](https://new.qq.com/rain/a/20260916A0APD700/))
• A loan fraud script list from June 2021 stated that scammers impersonated loan platform customer service and, under the pretext that bank card binding errors froze withdrawals, demanded borrowers provide a 3,800 yuan unfreezing fee, followed by chained fee collections under the bait of paying a 2,000 yuan processing fee to handle a 50,000 yuan loan. Legitimate banks do not charge such fees before disbursement, and the public should immediately report to the police when encountering upfront fee pitches. (Source: [https://zfw.changzhou.gov.cn/content/suitable/show?id=16704&catid=36878](https://zfw.changzhou.gov.cn/content/suitable/show?id=16704&catid=36878))

Official Stance

• On July 22, 2026, the website of the People's Government of Yunnan Province released anti-fraud tips pointing out that "hearing" may indeed be "false," reminding the public to be vigilant against AI synthetic voice scams and not to trust someone's identity based on voice alone.
• Anti-fraud centers across multiple regions densely issued warnings in 2026, alerting the public to new routines of AI voiceprint cloning loan fraud: strangers inducing the recording of digital voice videos, capturing voiceprints in 5 seconds to impersonate loans, and emphasizing that "loan platforms that charge money first are all scams."
• Chinese police repeatedly warned via official accounts in 2026 that any loan platform charging fees prior to disbursement under names like "risk control fees," "verification fees," or "data modeling fees" is in violation of regulations, and the public should immediately report to the police and preserve evidence.

How to Protect Yourself

• ✅ Never transfer money before loan disbursement: Refuse and terminate contact if any platform demands advance payment under any name before loan funds arrive. This is the core defense line.
• ✅ Verify platform financial qualifications: Check whether the lending institution holds financial licenses through the official websites of the CBIRC or local financial regulatory bureaus, rather than judging authenticity solely based on websites and customer service pitches.
• ✅ Refuse to provide biometric materials such as voice and face: Do not casually record voiceprints, read numbers aloud, or perform nodding/head-shaking videos to prevent biometric features from being cloned for subsequent fraud.
• ✅ Prioritize financing from banks or licensed consumer finance companies: Go through formal channels whenever possible when in financial need, maintain vigilance against packages like "instant AI approval" and "unsecured with no credit check," and immediately dial the 96110 anti-fraud hotline to report suspicious activity.