AI Batch Voice Robocalling Harassment & Scams: Impersonating Authorities / Acquaintances for Precision Fraud via Mass Dialing

Who Gets Targeted

Victims span all age groups, with individuals over 60 accounting for the highest proportion. This demographic has low awareness of AI technology and is easily terrified by deceptive scripts such as 'your child had an accident' or 'law enforcement investigation,' leading them to blind money transfers in a panic. E-commerce buyers and business owners aged 18 to 35 are primary targets for harassment, frequently manipulated by fake 'e-commerce customer service' or 'bank staff' scripts into leaking personal information and clicking phishing links. Newly graduated job seekers are easily lured by AI calls impersonating HR departments to pay training fees and deposits. Victims universally harbor natural trust in official identities and familiar voices, possess weak awareness of risks from unknown calls, and tend to lose judgment when faced with urgent scripts, thereby falling into scams.

骗局怎么运作

• Black and grey industry syndicates first purchase or crawl citizen personal information through illegal channels, including names, mobile numbers, kinship relationships, consumption records, and enterprise details, and purchase leaked voiceprint databases on the dark web to build data reserves for subsequent precision fraud. The core mechanism matches high-value targets via information asymmetry—such as filtering elderly people whose children live elsewhere, users with recent e-commerce consumption, or business owners with loan records as precision strike targets—to increase the success rate of scams.
• Syndicates purchase or self-deploy AI outbound calling systems configured with voiceprint cloning, real-person speech synthesis, and intelligent multi-turn conversation modules. They forge the voiceprint and customized script templates of the target's acquaintances or official institutions, and use carrier pass-through technology to tamper with caller IDs to display official numbers or friend-and-family numbers, evading carrier harassment-blocking systems and lowering the target's alertness.
• The system automatically makes batch calls to target phones, initially collecting the target's voiceprint characteristics via 'silent calls'. Pre-recorded scam or harassment scripts are played only after voiceprint matching is confirmed—such as a child impersonator saying 'Mom, I got into a car accident and urgently need surgery money,' or a customer service impersonator saying 'The product you bought has quality issues and requires a refund and compensation.' A single device can handle 1,000 to 1,500 outbound calls per day, far exceeding the efficiency of manual dialing.
• For hooked targets, the AI system guides them to add counterfeit customer service WeChat accounts, click phishing links, download fake apps, or directly induces money transfers. Certain high-value targets are transferred to human syndicate members for follow-up, using scripts such as 'secure account,' 'verifying funds,' or 'suspected illegal involvement requiring cooperation with an investigation' to swindle large sums of money. Once successful, they immediately transfer illicit funds through multi-tier accounts and destroy evidence.
• Syndicates also mark users who return calls or leak personal information as 'high-value targets,' subsequently pushing precision scripts for financial or training scams, or selling the users' mobile numbers and personal information to other grey industry syndicates for profit, forming secondary or multiple harvests to maximize the value extracted from victims.

红旗信号（看到这些快跑）

• 🚩 The caller ID shows an official number, but immediately upon answering, demands money transfers, bank card numbers, or verification codes; alternatively, the voice feels familiar but the speed and tone have subtle differences from usual, and the caller rushes you to make an immediate decision without allowing you to hang up or inform others.
• 🚩 Incoming unknown calls start with silence, static noise, or only a mechanical electronic tone before connecting to human or AI scripts. These are harassment or scam calls currently collecting voiceprints, and can be hung up immediately.
• 🚩 The caller claims to be customer service, law enforcement, or friends and family, but cannot answer private questions only known to the individual—such as failing to answer when counter-questioned with 'What photo did I send you last week?' or 'Where did we last eat together'—or requests downloading apps from unofficial app stores or clicking unfamiliar links.
• 🚩 Frequently receiving sales and scam calls from different numbers where the caller can accurately state your name, consumption records, relative information, and work details, indicating your personal information has been leaked and is being targeted by grey industry syndicates for precision fraud.
• 🚩 The caller exerts pressure using urgent scripts such as 'account frozen,' 'suspected illegal activity,' or 'family emergency,' demanding that you must absolutely not hang up or inform family and friends, and must immediately transfer money or complete verification operations. This is a classic scam pressure tactic.

真实案例

• In March 2026, Minhou police in Fujian reported an AI-cloned voiceprint scam case: A scam syndicate obtained information about the children of a retired teacher surnamed Jia through illegal channels, cloned their voiceprints, and called Jia claiming that 'an accident occurred requiring 5,000 yuan in surgery money urgently.' Jia was nearly tricked into transferring money before bank staff discovered it and reported it to the police. Police dismantled an AI outbound calling scam syndicate spanning 3 provinces, arrested 12 criminal suspects, seized 27 involved AI outbound calling devices, and recovered over 2.3 million yuan in involved funds.
• In July 2026, CCTV News exposed the chaos of AI outbound call harassment: A technology company deployed an AI outbound call system providing bulk telemarketing services to educational institutions and health supplement merchants, making up to 12,000 calls per day and evading user blocks by tampering with caller IDs. It had illegally acquired over 1.2 million pieces of citizen personal information, was investigated and penalized by regulatory authorities with an 800,000 yuan fine, and relevant responsible persons were subjected to administrative detention according to law.
• In May 2026, an enterprise owner surnamed Yi in Shanghai received an AI call impersonating 'bank customer service.' The caller accurately stated Yi's corporate loan information, induced Yi to click a link to download a fake 'bank app,' and transferred 1.2 million yuan from the account after filling in the bank card number and verification code. After initiating an investigation, police discovered the syndicate specifically targeted users with corporate loans, having committed 17 crimes involving over 6 million yuan.
• In April 2026, Xinhua News Agency reported that the Jinchuan Branch of the Jinchuan City Public Security Bureau in Gansu Province cracked an AI voice impersonation official customer service scam case: Outlaws used AI speech synthesis technology to impersonate official customer service of short-video platforms, inducing elderly people to download apps and transfer funds from their cards. The Jinchuan Branch task force arrested 8 criminal suspects involving over 20 victims, mostly elderly, with a total amount exceeding 3 million yuan. (Source: [https://www.news.cn/legal/20260428/ccc23e82b9574c7ca1a641d0d10bbf26/c.html](https://www.news.cn/legal/20260428/ccc23e82b9574c7ca1a641d0d10bbf26/c.html))
• In April 2026, citing police reports from various regions, Xinhua News Agency reported that an 82-year-old living alone was swindled out of 830,000 yuan in retirement savings by 'fake customer service,' and another sexagenarian fell victim to a 'fake customer service' scam resulting in over 10,000 yuan of bank card balance being transferred. Police in Gansu, Jiangsu, Shandong, and other regions reported multiple cases of elderly people being scammed by fake official customer service using AI speech synthesis. (Source: [https://www.news.cn/legal/20260428/ccc23e82b9574c7ca1a641d0d10bbf26/c.html](https://www.news.cn/legal/20260428/ccc23e82b9574c7ca1a641d0d10bbf26/c.html))

Official Stance

• In April 2026, the National Anti-Fraud Center issued an 'Emergency Warning Regarding AI Voice Scams,' reminding the public that if unknown incoming calls involve money transfers or requesting verification codes, they must verify the caller's identity through original contact methods, and remain alert to new scam methods such as 'silent calls' and 'AI-cloned acquaintance voices.'
• In July 2026, the Ministry of Industry and Information Technology and the Ministry of Public Security jointly issued the 'Notice on Launching Special Rectification Actions against AI Outbound Harassment Calls,' explicitly demanding a comprehensive cleanup of illegal AI outbound calling systems, a severe crackdown on grey industry syndicates using AI technology to commit telephone harassment and scams, and achieving a reduction of over 90% in AI outbound harassment call volume by the end of 2026.
• In August 2026, the Supreme People's Court released 'Typical Cases of Telecommunication Network Fraud Crimes,' explicitly classifying the acts of using AI voiceprint cloning and batch outbound calling to commit fraud as the crime of telecommunication fraud, punishable by up to life imprisonment and fines, with principal offenders of organized syndicates punished severely according to law.

How to Protect Yourself

• ✅ Enable anti-harassment features on your mobile phone, download the National Anti-Fraud Center app, and enable 'Incoming Call Warning' permissions. If unknown incoming calls involve money transfers or request verification codes, be sure to hang up and verify the caller's identity using contact information from your original address book, rather than blindly trusting the caller ID number.
• ✅ When receiving AI calls suspected to be from acquaintances, proactively ask private information known only to both parties (such as the location of a recent meal together or common friends' nicknames) to verify identity. Do not be frightened by urgent scripts like 'something happened' or 'suspected illegal involvement,' and do not follow the other party's demands to 'keep it confidential' or 'not hang up.'
• ✅ Properly safeguard personal ID numbers, mobile numbers, bank card numbers, voiceprints, and other biometric information. Do not casually input voiceprint information on unfamiliar platforms or apps, and directly hang up when receiving silent calls to avoid having voiceprints collected for clone scams.
• ✅ If accidentally scammed, call 110 to report it immediately, while preserving evidence such as call logs, chat histories, and transfer vouchers. Cooperate with police investigations, contact banks in time to freeze accounts, and minimize losses to the greatest extent possible.