To a Shanghai man, the woman calling herself “Tingting” seemed like someone with whom he could build a future.
Discovered on a short-video platform, her profile featured an attractive image of a young woman. She presented herself as an orphan raised by a devoted aunt, and the two quickly moved their conversations to WeChat, the most used social media in China. Their online relationship blossomed, with daily exchanges that eventually turned to marriage plans. She shared glimpses of her supposed everyday life, including photographs of meals, and appeared attentive and affectionate.
He believed he was preparing for marriage.
Instead, the woman he hoped to marry did not exist. Her identity was part of an elaborate romance scam built around images generated with artificial intelligence (AI) and a carefully choreographed seven-day messaging script. By the time the deception unraveled, the man, surnamed Liu, had transferred more than 180,000 yuan (about $25,000) to a criminal syndicate—and even took out a small loan to keep the money flowing.
The case, uncovered by police in Shanghai’s Minhang District, involved 15 male victims across China who collectively lost more than 1.8 million yuan (about $250,000), according to a CCTV News report on Nov. 2.
A 7-Day Blueprint
What made the operation particularly striking was its assembly-line approach to romance.
Investigators found that the criminal group employed dedicated chat operators, who followed a standardized “seven-day chat manual” to cultivate relationships with male targets.
The operators were instructed to send AI-processed lifestyle photographs, including pictures of meals, to make their fictional identities appear more convincing. On the second day, they were to introduce complaints about working overtime, creating opportunities for sympathy and emotional connection. By the third day, they were encouraged to switch from scripted responses to manual typing, making conversations feel less mechanical and more spontaneous.
Once a relationship had warmed up, the script introduced a seemingly modest financial request.
The woman would claim that her boss was harassing her and that she intended to resign and open a business. She would then ask her online boyfriend to send a few hundred yuan for flower baskets to celebrate the planned opening.
The flowers, however, were not the point. The request was designed to identify men who could be persuaded to send money.
Those who agreed were considered promising targets and passed along for further manipulation. Those who refused to buy flowers or gifts were marked as having no potential in the operators’ contact lists and dropped from further pursuit.
For the syndicate, even a small payment was enough to determine whether a romantic prospect was worth pursuing.
A Fabricated Family Crisis
Liu passed the test.
After roughly a month of online courtship, Tingting suddenly told him that the aunt who had raised her had been diagnosed with a brain tumor.
To make the emergency appear genuine, she sent him images of medical diagnostic documents concerning her aunt, along with a photograph of her own identity card. Convinced by the evidence, Liu initially transferred 40,000 yuan (about $6,000) to help cover the supposed medical expenses.
Over the next two weeks, he sent more money, eventually exceeding 180,000 yuan. He even resorted to small online loans to continue helping the woman he believed would become his wife.
Behind the scenes, the supposed family tragedy was being manufactured at remarkably little cost.
Police discovered that the syndicate had fabricated the supporting materials, including medical documents, identity card images, and photographs purportedly showing the aunt in a hospital bed. Producing a forged identity card cost just 25 yuan ($3.75), while image-editing work cost approximately 10 yuan ($1.50) per picture.
The contrast was stark: A handful of inexpensive fabricated images helped persuade one victim to surrender more than 180,000 yuan.
The Video Call Test
The deception finally began to unravel when Liu could no longer afford to send money.
Wanting to verify his girlfriend’s story, he insisted on a real-time video call to see the hospitalized aunt. Instead of agreeing, Tingting provided him with a hospital address.
The response raised Liu’s suspicions, prompting him to report the matter to police.
An investigation by Shanghai police found that Liu was not alone. The same fictional woman and the same fraudulent operation had ensnared men in different parts of China.
Police eventually arrested five members of the syndicate. Investigators determined that the group had defrauded 15 victims of more than 1.8 million yuan ($250,000) in total, with the largest individual loss exceeding 380,000 yuan ($57,000).
The defendants were subsequently sentenced for different offenses. The longest prison term was 11 years, and fines were also imposed.
AI Meets Old-Fashioned Deception
The Shanghai case illustrates how AI technology can make traditional romance scams more convincing without requiring sophisticated technology.
In a February 2024 report, China’s state-run Xinhua News Agency warned that advances in AI have enabled criminals to superimpose one person’s face and voice onto another in real-time video, creating highly realistic synthetic content for new forms of online fraud.
But Liu’s experience shows that a convincing real-time deepfake may not even be necessary. AI-processed photographs, fabricated documents, human-operated conversations, and a carefully timed emotional appeal were enough to sustain the illusion, turning a man’s hopes for love and marriage into an opportunity for financial exploitation.
Zhao Bin contributed to this report.
