Make 2,000 yuan a day, AI "sons" defraud the middle-aged and elderly in batches.
Recently, a child named "Xiao Hai" has gone viral on short video platforms.
In the footage, this teenager who looks to be around ten years old is always crying, telling "aunt moms" his miserable life story: his mother left home, his father is hospitalized, he is mocked by classmates, and he relies on livestream sales to subsidize the family on a daily basis. A large number of middle-aged and elderly women in the comment section leave comforting messages while placing orders to support him.
But Xiao Hai is actually an AI digital human. If you search for keywords on short video platforms, you will find more than a dozen accounts with different IPs sharing the exact same face, and all videos are uniformly marked "Content generated by AI". Up to now, the view count of the Douyin topic #Xiao Hai has exceeded 600 million times.
What does the AI digital human Xiao Hai rely on to attract traffic and realize monetization? How is this set of AI-driven pity-playing product sales model turned into an assembly line operation? Today we are going to talk about these topics.
01 From "Brother Jianguo" to "Son Xiao Hai", AI Digital Humans Harvest Retired Aunts
Many people may not have expected that AI digital human product sales targeting the silver-haired group has long developed into a family-style business. The pity-playing sales model of AI Xiao Hai is already the iterated and upgraded 2.0 version, and the first generation version was those AI overbearing presidents that once went viral all over the internet.
In early 2025, a large number of AI middle-aged men emerged in batches on short video platforms. Some wear shirts paired with jackets, with flowers, plants and blue sky as the background, presenting a government-affiliated style; some are dressed in formal suits, standing in villas and conference rooms; there are also retired veterans appearing in aircraft cabins and under the national flag. Their names are also very down-to-earth, which are most familiar to the middle-aged and elderly people such as "Jianhua", "Jianguo" and "Dongdong". Their personal settings are also unified: single, well-off, and all want to find a life partner.
AI overbearing presidents who speak cheesy love lines have attracted a large number of middle-aged and elderly female fans
The content templates are highly standardized, with dozens of seconds of short videos or static images and texts, paired with AI-synthesized voices. The lines are basically the same over and over again: singing love songs, saying cheesy love lines, and confessing to the camera: "Although years have left traces on your face, it does not affect your charm at all"; some even kneel down to propose directly: "Can I ask my wife to marry this old man of mine?", and even promise: "I have a net worth of tens of millions, and I will transfer a large furnished villa to you".
The same AI image and the same set of lines are copied to different platforms and dozens of accounts. Many users keep swiping across the same "man", but when they click into the homepage, it is a completely different account.
At that time, the most popular Wang Jianguo was also called "Brother Jianguo" by the retired aunts. His personal setting was carefully polished: in his forties, working in a government-affiliated institution, divorced and raising his child alone, affectionate, family-oriented, and leading a hard life. The videos have no complicated plots, just monologues to the camera, starting with "dear sister", interspersed with various grievances in life, unsatisfactory work, raising children alone, and losing sleep at night. Many accounts can update several videos a day, with extremely high content repetition.
After the traffic logic was verified, the monetization chain followed immediately. The videos of Brother Jianguo began to sell products: "Sister, this is the good product I specially found for you. You are usually reluctant to buy it, and I feel bad for you." It guides fans to click on the homepage showcase to place orders. The products are mainly health supplements, skin care creams and food, with prices ranging from 19.9 yuan to 99.9 yuan, and some product links have monthly sales of over 10,000 units.
The AI digital human Wang Jianguo even "brought his whole family along": his son Xiao Hai and daughter Xiaoya also began to appear on camera, register accounts, and carry out IP matrix operation.
AI Xiao Hai has a handsome and well-behaved appearance, lives with his sick father, comes from a single-parent family and lacks maternal love. His core positioning is family companionship, he often shouts "aunt moms" to the camera, asks about their daily lives from time to time, and sings and dances to make fans happy occasionally. The corresponding pity-triggering hooks have also been adjusted: almost winning a competition, falling behind in rankings, being bullied by classmates at school, after sufficient emotional preparation, when Xiao Hai cries emotionally, he will take out cooking oil and condiments from his arms, hold them up to the camera, and ask the aunt moms to place an order. All the sad stories eventually lead to the shopping cart link below the video.
AI Xiao Hai accounts with highly repetitive content
Following the same account replication logic as "Jianguo", a number of accounts such as @Sensible Xiao Hai, @Haihai Shares Daily Life, @XiaoHaiXiaoHaiXiaoHai were launched one after another, accumulating 10,000 followers, and the total view count of related topics exceeded 600 million times.
From AI Wang Jianguo to AI Xiao Hai, these videos have rough picture quality, stiff expressions, and often mismatched lip sync, and are marked with "Content generated by AI". Young people can see through them at a glance, but it is very difficult for the elderly to tell the difference.
The core fan portraits of the two generations of IPs are highly overlapped, precisely targeting middle-aged and elderly women in their sixties and seventies, covering more than 20 provinces across the country. If you open the comment section, there are a large number of voice-to-text messages, and after dialect recognition, the word order is messy with many grammatical errors.
Operation Society's analysis believes that from the AI middle-aged overbearing presidents under the guise of blind date and romance, to the single-parent teenager Xiao Hai who plays the pity card to ask for help, the underlying logic is consistent: precisely targeting the emotional gap of the silver-haired group, building trust through emotional bonding, and then completing follower growth and product sales monetization.
Even after being exposed by the media many times and regulated by platforms, this business has not disappeared.
Operation Society observed that the account @Brother Jianguo has 61,000 followers on Douyin. Although it has been banned from new follows, it is still updating recently. There are also a number of shell-changing accounts on the platform such as @Walk Hand in Hand (Guoguo), @Good Luck Always, @Daily Life of Brother Guo, @Precious Brother Jianguo, which keep updating with the #Jianguo and #Sister Bao tags, with the number of followers ranging from 2,000 to 30,000. In addition, more than a dozen derivative live streaming rooms are broadcasting at the same time: the background displays the image of AI Wang Jianguo, with real human assistants doing voice interaction to guide fans to send gifts, and the number of likes per live stream stays stable at around 20,000.
Brother Jianguo's shell-changing live stream
02 Same Script with Different Shells: How Does the AI Digital Human Gray Industry Become an Assembly Line Business?
Platform supervision has been tightening. The special action "Qinglang · Rectify the Chaos of False Personal Settings on Short Video Platforms" launched in September 2026 lists "pretending to be vulnerable groups" as the top priority of rectification, and all behaviors such as account leasing and selling, proxy identity verification, and proxy enterprise certification are included in the disposal scope. Now the customer acquisition cost of registering new accounts is getting higher day by day. But Wang Jianguo and Xiao Hai still keep appearing under new accounts.
So, how does this AI digital human product sales model focused on emotional hunting achieve standardized mass production? And why does it keep popping up despite repeated bans?
20 Yuan to Copy a Face, Using Standardized Personal Settings to Hunt Middle-aged and Elderly People
AI digital human is not a new concept now, and AI content generation is widely used in e-commerce product sales and grass-planting videos. Compared with many e-commerce merchants, the content quality of Wang Jianguo and Xiao Hai is actually not high, with stiff expressions, often mismatched lip sync and rough picture quality, and the technical threshold is extremely low.
But its real advantage lies in a set of infinitely replicable "personal setting + script" templates: AI is only responsible for providing a face, the personal setting is responsible for providing emotional connection, and the relationship itself is the reason for users to pay.
The scripts are so standardized that no creativity is needed. The fixed structure is "elderly addressing + daily troubles + shopping link asking for help". The core personal setting remains unchanged, and the scenes and lines change according to product categories. When selling face cream, the beauty salon scene is used; when selling socks, the temperature difference between day and night is mentioned. The core tone is always the tone of younger generations chatting with elders.
This template precisely hits the emotional gap of middle-aged and elderly women in their sixties and seventies: their children are busy with work, and they lack companionship and people to talk to in daily life. A "younger generation" who is willing to chat with them every day and cares about their trivial matters can easily build emotional trust. Wang Jianguo takes the "reliable old brother" route, chatting about daily life to the camera, making the aunts develop similar emotional dependence; Xiao Hai focuses on the "sensible son" personal setting, greeting them with care, and guiding them to place orders with pity-playing lines such as "failed the exam" and "being laughed at by classmates for wearing old clothes".
The content cost is also very low. A digital human video within 30 seconds with perfectly aligned lip sync costs as low as 20 yuan per piece. An industry insider revealed that the teams behind those active accounts of Jianhua and Jianguo have long debugged the mature process, "which can easily realize mass production".
A media reporter followed the public tutorials online to practice, and the whole process from registering an account, importing the digital human image to posting the first video only took more than ten minutes. 24 hours after posting, the total view count on the two platforms exceeded 4,000, and more than 60 precise followers were gained.
Dozens of Accounts Running at the Same Time, Minimizing the Account Ban Cost to the Extreme
The short life cycle of a single account and account ban due to violations are the default premise of this track. The core purpose of deploying a large number of accounts in batches is to use quantity to hedge risks, and dilute the account ban cost to a negligible level. A behind-the-scenes operation team revealed that they usually operate dozens of video accounts at the same time, "these accounts will 100% violate the rules, and one account can survive for a maximum of 5 days", but they can earn at least 2,000 yuan a day relying on product sales commissions. Rough estimation shows that a single account can generate about 10,000 yuan of revenue during its life cycle. When dozens of accounts run in parallel, the revenue of a single account accounts for less than 2% of the total revenue.
With batch accounts and matrix operation, the risks are also diluted. If one account is banned, there are still a large number of others left. The accounts are registered under different real-name entities in different provinces. Even if they are investigated, the investigators can only find the outermost account leaser or trainee, and it is very difficult to find the behind-the-scenes operator. There is no need to bet on a single account to go viral, and once an account performs well, more traffic will be invested in it to adopt the horse-racing mechanism. Once a face is seen through, the same script can be used to register new accounts by changing the face, name and identity, with very low restart cost.
But there is still an unavoidable threshold: account real-name authentication and face verification. One person cannot collect so many identities at all. This is also the reason why the supporting business of "recruiting trainees" is booming. On the surface, it sells courses to teach you how to make money by running AI accounts, but in fact it collects the real-name information of the trainees. Some operation teams have hundreds of trainees under them. Trainees pay for the templates, register accounts with their own identities, and take full responsibility for all account violations. The operators are only responsible for operation guidance and share the commission. A simpler path is to directly lease and buy accounts, which even saves the process of recruiting trainees.
Earn Commissions from Public Domain Product Sales, Earn Tuition Fees from Private Domain Course Sales
Once the accounts get traffic, the monetization of this industrial chain is divided into two layers.
The first layer of revenue comes from public domain showcase product sales, earning the information gap profit from the elderly. The products in the showcase are mostly food and daily chemical products priced from dozens to 200 yuan, and the prices are generally much higher than those in formal channels: perilla seed oil of the same specification sells for 60 to 70 yuan in supermarkets, but the link is priced at 99 yuan; donkey-hide gelatin of the same brand sells for more than 60 yuan in other channels, but it also sells for 99 yuan here. The bulk of the price difference is the commission for the product sellers. According to industry statistics, the product commission rate of short video accounts targeting the elderly track is generally 50%-70%, much higher than other categories.
The real harvesting comes later. After the elderly leave messages in the comment section, a "private small account of the younger brother" will soon add them on WeChat. After being approved, they will be pulled into a group to grab red envelopes, and then guided to third-party private domain live streaming rooms without platform supervision. In the live streaming room, the host sometimes acts as a caring person who bargains with the factory for the sisters, sometimes acts as a health expert. After narrowing the distance through interactive Q&A, they sell health supplements with a cost of only more than ten yuan that are claimed to be able to regulate chronic diseases, whose profit is much higher than public domain product sales.
But compared with the product sales commission of dozens of yuan per order, the really steady profit comes from course sales. One set of script templates can be sold to unlimited number of people. Even if all accounts are banned, the tuition fees have already been collected. Some teams launch a "one-on-one accompanying operation" service priced at 998 yuan, which includes video production, account operation, and supplier connection, promising to "realize monetization in about a week after account registration".
To sum up, the AI digital human product sales model that hunts the middle-aged and elderly has formed a complete industrial chain division of labor: people who write scripts, people who make images, people who edit and distribute content, people who operate showcases, and people who divert traffic to private domains. Every link has dedicated personnel in charge, and the operation is highly mature.
03 Conclusion
The continuous iteration of AIGC large models is pushing AI digital humans from concept to large-scale implementation. The texture of generated content is constantly improving, and the production threshold continues to drop. Digital humans have already penetrated into various scenarios of short video product sales and content creation. More and more users begin to browse AIGC content. According to QuestMobile data, by May 2026, among the interest preferences of users in the sinking market, the active proportion of AIGC has reached 47.3%, an increase of 28.3 percentage points in one year.
The commercial monetization model around AI digital humans has also been fully verified, which