AI Online Dating Multi-Role Scam: Fraudsters Use AI Clones to Batch-Generate Virtual Friends and Family Circles to Sequentially Harvest Victims' Savings

Who Gets Targeted

Victims are predominantly single individuals aged 25 to 50, including urban white-collar workers, divorced middle-aged people, middle-aged and elderly single men living alone, and overseas Chinese workers. Generally emotionally lonely, with narrow social circles and weak awareness of AI technology, they easily project their desire for a perfect partner onto meticulously crafted virtual personas. Their psychological vulnerabilities center around four dimensions: emotional starvation, impulsive trust, concern for social face, and the sunk cost fallacy. Once emotionally invested, victims are reluctant to face the reality of being scammed, sometimes even actively defending the scammers despite dissuasion from friends and family. Some victims collapse and wake up only after losing their life savings, yet choose not to report it out of shame, making deep-seated loneliness render them targets for repeated hunting.

骗局怎么运作

• Step 1: Account Nurturing and AI Persona Building. Fraudsters register accounts on social or matchmaking platforms, using AI to generate high-quality lifestyle photos and short videos to portray wealthy, handsome men or beautiful women. Simultaneously, they use voice cloning technology to batch-generate voice messages, allowing victims to build trust across text, image, and voice dimensions simultaneously—far surpassing the thin image of traditional pig-butchering scams that rely solely on stolen photos.
• Step 2: Precise Screening and Multi-Channel Casting. Operating hundreds of clone accounts across matchmaking platforms, short-video comment sections, and interest communities simultaneously, the gang casts lures, utilizing AI to automatically match and filter targets with fast response times and high emotional needs, steering high-intent users to private domains like WeChat for one-on-one engagement, greatly enhancing traffic conversion efficiency and screening precision.
• Step 3: Sequential Role Implantation During the Emotional Nurturing Stage. After the main scammer establishes a romantic relationship with the victim, they immediately use the same AI tool to play multiple roles such as a best friend, mother, colleague, and doctor who take turns appearing. Combined with posting Moments, voice calls, and virtual red packets to create a complete social ecosystem, this makes the victim mistakenly believe they have integrated into the other person's social circle, expanding trust from a single person to the entire fake social network.
• Step 4: Fabricating Crisis Events to Execute Sequential Emptying. The scammers arrange for characters to successively encounter emergencies such as illness and hospitalization, project capital turnover difficulties, and customs duty payments, requesting advance payment for medical expenses under the guise of family illness, hinting at insider information through investment opportunities, and drawing the victim into a fake investment platform under the pretext of a shared future. Each amount increases progressively, utilizing the victim's sunk cost fallacy to continuously add funds until all savings are emptied.
• Step 5: AI-Assisted Multi-Line Parallelism and Counter-Reconnaissance. The gang uses AI script libraries to automatically generate daily greetings, emotional response templates, and standard answers to doubts, allowing a single person to control hundreds of conversation windows simultaneously. When a victim becomes suspicious, they switch to other roles to comfort them or manufacture accidents to interrupt contact, cooperating with the backend control of the virtual investment platform to let victims see book profits while remaining unable to withdraw them. Upon completing the final harvest, they destroy the accounts and transfer the funds.

红旗信号（看到这些快跑）

• 🚩 The other party's photos are exquisite yet stylistically uniform like AI generation, and they refuse real-time video calls or disconnect after only a few seconds each time, claiming poor signal or work confidentiality, unable to provide consecutive self-proving footage.
• 🚩 Mentioning investments, blockchain, and insider guaranteed-profit opportunities early in the relationship, claiming to have achieved financial freedom through investment, and frequently sending fake trading platform profit screenshots to create an atmosphere where friends and family are also co-investing.
• 🚩 Surrounding friends, colleagues, and other roles suddenly appear intensively and all add the victim as a friend, yet the content of these people's Moments is identical, interaction times overlap, and their profile pictures and voice styles are highly similar to the main scammer.
• 🚩 Requesting to migrate conversations from the platform to WeChat or encrypted messaging apps under the excuse of platform throttling or convenience of contact, subsequently gradually exerting emotional pressure in the private domain and isolating the victim's original social support system.
• 🚩 Requesting transfers to personal accounts rather than corporate accounts under the pretext of illness, hospitalization, project turnover, or customs tax payments, where the recipient does not match the claimed identity, while urging urgency, emphasizing confidentiality, and forbidding discussion with family.

真实案例

• According to Zhejiang police briefings and NetEase News reports, a woman in Zhejiang used WeChat clones over the span of 3 years to play multiple roles including a mother and best friend, defrauding a male netizen of a total of 450,000 RMB. After squandering the money, feeling guilty, she asked AI what she should do, and being persuaded by AI, she went to the local police station to surrender herself. She has now been criminally detained, and the case details have been publicly disclosed by official media. (Source: [https://www.163.com/dy/article/L362JHKU051492LM.html](https://www.163.com/dy/article/L362JHKU051492LM.html))
• According to HK01 reports, a woman in Shenzhen fell victim to a scam ring posing as someone with a Hong Kong background in an online romance, ultimately resulting in tragic bankruptcy. The gang involved over 200 victims, and the fraudsters utilized meticulously packaged identities and emotional scripts to cultivate long-term relationships on social platforms before introducing fake investment. The case has been intervened and investigated by law enforcement agencies. (Source: [https://www.hk01.com/%E5%A4%A7%E5%9C%8B%E5%B0%8F%E4%BA%8B/60375474/%E6%B7%B1%E5%9C%B3%E5%A5%B3%E7%B6%B2%E6%88%80%E5%A2%AE%E6%AE%BA%E8%B1%AC%E7%9B%A4%E6%85%98%E7%A0%B4%E7%94%A2-%E8%A9%90%E9%A8%99%E5%9C%98%E5%A4%A5%E5%81%87%E6%89%AE%E9%A6%99%E6%B8%AF%E8%83%8C%E6%99%AF-%E9%80%BE200%E4%BA%BA%E5%8F%97%E9%A8%99](https://www.hk01.com/%E5%A4%A7%E5%9C%8B%E5%B0%8F%E4%BA%8B/60375474/%E6%B7%B1%E5%9C%B3%E5%A5%B3%E7%B6%B2%E6%88%80%E5%A2%AE%E6%AE%BA%E8%B1%AC%E7%9B%A4%E6%85%98%E7%A0%B4%E7%94%A2-%E8%A9%90%E9%A8%99%E5%9C%98%E5%A4%A5%E5%81%87%E6%89%AE%E9%A6%99%E6%B8%AF%E8%83%8C%E6%99%AF-%E9%80%BE200%E4%BA%BA%E5%8F%97%E9%A8%99))
• According to United Daily News reports, Taiwan police and prosecutors cracked an AI voice-cloning romance scam case, arresting 6 public relations hostesses who used AI voice cloning technology to disguise themselves as women to date male clients online, earning about 100,000 New Taiwan Dollars per month. Victims were induced to continuously tip and transfer funds, and the case has been investigated and publicly disclosed by prosecutors and police. (Source: [https://udn.com/news/story/124490/9604161](https://udn.com/news/story/124490/9604161))
• In July 2026, according to Liberty Times reports, prosecutors from the Taipei District Prosecutors Office directed police to investigate an AI voice-cloning romance scam operation center. The fraud ring used AI voice-cloning software to disguise female vocal tones and arranged female hostesses to date netizens, tricking them into purchasing mobile phones, gold, and high-priced courses. Financial losses were crudely estimated at several million NTD, and police and prosecutors detained 12 people. (Source: [https://news.ltn.com.tw/news/society/breakingnews/5491936](https://news.ltn.com.tw/news/society/breakingnews/5491936))
• In June 2025, according to Rule of Law Daily reports, Quanzhou Nan'an police in Fujian cracked a matchmaking and dating fake investment and wealth management scam case. The ringleaders established an overseas fraud stronghold, and the gang packaged personas to win trust before inducing investments. The case involved an amount of 180 million RMB and was jointly listed for supervision. (Source: [https://hxfzw.net/anjian/3596.html](https://hxfzw.net/anjian/3596.html))

Official Stance

• The Criminal Investigation Bureau of the Ministry of Public Security issued an anti-fraud warning in 2024, pointing out that pig-butchering scams are leveraging AI technology to upgrade script and image camouflage capabilities, reminding the public to be vigilant against transfer requests from online lovers under the guise of investment and borrowing, and emphasizing that any request for money transfers before meeting in person constitutes a high-risk signal.
• The National Anti-Fraud Center App pushed multiple early warnings in 2024, reminding users to be wary of the application of AI face-swapping and AI voice-cloning technologies in online dating scams, suggesting that when receiving video or voice calls, users should ask the other party to perform designated actions to verify authenticity, and enable the incoming call warning function.
• The Supreme People's Procuratorate reported at a 2024 anti-fraud thematic press conference that cases utilizing generative AI to batch-produce fake identities for emotional fraud are on the rise, emphasizing that crackdowns on AI-assisted fraud will be intensified, while reminding the public not to blindly trust investment advice from internet lovers.

How to Protect Yourself

• ✅ Before establishing a romantic relationship online, insist on requiring the other party to conduct at least one complete real-time video call, and ask them to perform designated actions during the call such as raising a hand or turning their head to rule out the possibility of pre-recorded videos and AI face-swapping.
• ✅ Install and enable the incoming call warning and text message detection functions of the National Anti-Fraud Center App. When encountering links or QR codes where individuals claiming to be relatives, friends, or lovers request money transfers, use the app's built-in detection tools to scan and verify, while maintaining maximum vigilance toward unfamiliar receiving accounts.
• ✅ Once the other party mentions investments, insider information, guaranteed-profit opportunities, or requests money transfers to advance medical expenses, security deposits, etc., immediately stop the conversation and call the 96110 anti-fraud hotline for consultation, saving screenshots of the chat history as evidence for subsequent reporting.
• ✅ Proactively disclose the existence and basic background of the online romantic interest to friends and family before investing emotionally, listening to third-party objective judgments. Once red flags such as the other party refusing video calls, avoiding meetings, or urging money transfers are discovered, immediately terminate the relationship and preserve all chat and transfer records.