AI-generated "tyrant president" targets middle-aged and elderly women
"Sister, I understand all the hardships you've endured. From now on, with me around, I will never let you suffer any grievance again."
At 3 a.m., the 63-year-old Aunt Chen's phone screen was still lit. On the screen was an elegant man in a sharp suit, speaking affectionate words directly to the camera. Aunt Chen tapped the comment section and carefully typed, "Thank you for your concern, little brother. Sister loves you." The man in the video was actually an AI-generated digital human, and in the comment section of this video, there were thousands of people like Aunt Chen interacting with him.
The outside world rarely sees the full picture behind this boom: The root cause of the elderly being addicted to AI emotional videos lies in the unplaced emotional loneliness in reality. The assembly-line mass production of content and the continuous feeding of platform algorithms have jointly woven a gentle emotional hunting net targeting the silver-haired group. The simple mockery of "love obsession" from others cannot hide the essence of this full-chain exploitation.
01. Low-cost Mass Production: How the Gray Industry Chain Operates
In the traffic world of short video platforms, "AI Overbearing President" is becoming a highly profitable business. From the "fake Jin Dong" incident to today's "AI Overbearing President", the exploitation targeting the emotional needs of the elderly has never stopped, but this time the sickle is faster, harsher, and more hidden. With the support of AI, creating hundreds of affectionate "overbearing presidents" a day is no longer a fantasy, but has evolved into a low-cost gray assembly line.
The picture shows different types of overbearing presidents generated by AI
The production of traditional emotional short videos relies on real actors, venue rental, and post-editing, and the comprehensive cost of a single video often reaches hundreds or even thousands of yuan. However, the mass production cost of a single "AI Overbearing President" video only costs a few yuan. According to Xinhua News Agency, "Due to the extremely low threshold of production cost, it takes at most 5 minutes to make a digital human video, and a single practitioner can operate dozens or even hundreds of accounts at the same time." This means that operators do not pursue making a lot of money from a single video, but spread the traffic through massive videos. Even if the income of a single video is only a few cents, the superposition of dozens of accounts and tens of thousands of videos can bring in thousands of yuan a day or even more.
When the marginal cost of a commodity approaches zero, no matter how low the price is, as long as the scale is large enough, the profit will show compound growth. The violent myth of "earning ten thousand yuan a day" in the gray training chain of AI emotional videos is essentially the inevitability of large-scale production with marginal cost approaching zero. Without assembly-line mass production, there will be no profit, and it can only be eliminated by the market.
From the perspective of the digital human image generation link, the development cost of traditional virtual images is high, but with the support of large AI models, the production threshold has been greatly reduced. According to The Paper, operators in the gray industry chain often only invest 9.9 yuan to 20 yuan to obtain a digital human model from the platform, and then use voice cloning for dubbing to produce in batches using it as a master version.
In the content production link, AI has realized one-stop generation from copywriting, dubbing to finished video, and the comprehensive cost of a single short video has generally been compressed to less than 10 yuan. Copywriting is almost zero-cost, and dubbing is also cheap—the open-source model IndexTTS2 is completely free, and it can be reused infinitely after one cloning. Video generation is the main expense, but the price has also dropped significantly—calculated based on the 659-yuan annual membership of the basic package of Jimeng AI, the cost of a single 15-second video is only about 2.5 yuan.
In terms of account operation, AI tools are also "reducing costs". With the help of matrix management tools like Yixiaoer, operators can manage 20 to 30 accounts at the same time by spending a few hundred yuan a year, and scheduled publishing, batch distribution, and automatic replies can all be handed over to AI tools. According to the *China Online Audiovisual Development Research Report (2026)*, the number of AI-generated audio and video content in 2025 exceeded 2 billion, a 14-fold increase over the previous year. These massive videos are continuously poured into the platform's traffic pool. Even if a single video only earns a few cents, the sum of thousands or tens of thousands of videos will bring considerable income.
With this low-cost, high-output assembly line, gray industry chains like "AI Overbearing President" can operate completely. From the upstream technology and material supply, to the midstream mass production, and then to the downstream traffic monetization, each link has a clear division of labor. And the only task downstream is to put the mass-produced "AI Overbearing President" content on short video platforms for monetization.
This monetization system can be roughly divided into three layers: the first layer is platform traffic sharing, where operators rely on mass replication to accumulate traffic and achieve considerable daily income. The second layer is window shopping with goods. After the number of fans accumulates to a certain level, the account starts to hang product links.
According to Xinhua News Agency, a new account with only 4 "AI Overbearing President" videos, each of which has about 1,000 likes, has already sold 33 products in its window, with a high customer unit price and a profit margin as high as dozens of times. The third layer is membership services and induced transfers. Operators set up fan groups to charge fees on a monthly or annual basis, divert fans to private domains to earn commissions, and even fabricate scripts such as "being sick and hospitalized" or "meeting in real life after marriage" to defraud large expenditures.
From the "small profits but quick turnover" of traffic sharing, to the "premium exploitation" of earning the price difference by selling goods, and then to the "single-point blasting" of inducing transfers, the three layers of monetization progress step by step, converting the emotional needs of the elderly into real money.
02. Emotional Vacuum: Why the Elderly Are Trapped
However, only low-cost content supply is not enough to form a "hunting". What really makes "AI Overbearing President" videos rampant is the huge emotional demand market behind them.
On the Internet, you can often see complaints from young people: grandma cooks a whole table of dishes in the middle of the night waiting for the "overbearing president" to pick her up; grandpa likes and leaves messages on AI videos like falling in love online. The original intention of these complaints is to worry that the elderly will be deceived, but few people ask: What exactly do the elderly want?
The picture shows an elderly person writing a love letter to the AI Overbearing President
This is actually the era predicament of a whole generation of elderly people. A survey by the Ministry of Civil Affairs shows that empty-nest elderly in urban and rural areas in China account for 59.7%, with a total number exceeding 180 million, among whom more than 38 million are completely living alone. Under the wave of urbanization, it has become the norm for children to study in different places and work across provinces. More and more elderly people are living a life of "locking the door when going out and turning on a single lamp when entering the house".
The fifth sample survey on the living conditions of the elderly in urban and rural China shows that 23.76% of the elderly have varying degrees of loneliness, and the loneliness problem of empty-nest and elderly living alone is particularly prominent. These elderly people have been in a state of lack of emotional communication for a long time. Once they encounter the overwhelming "sugar-coated bullets" of AI emotional videos, it is easy for them to sink.
Research from the Faculty of Psychology of Beijing Normal University points out that the core motivation for the elderly to watch short videos is that their emotional needs in reality are not met—they try to seek emotional connection and fill emotional gaps in this way. Tang Dan, a professor at the Institute of Gerontology of Renmin University of China, also reminds: Public opinion should not simply label the elderly's Internet behavior as "being deceived" or "addicted". We need to abandon stereotypes and see the real emotional predicament behind it.
Many children understand filial piety as providing material satisfaction, but few people realize that what the elderly really crave is that someone cares about their emotions. The breakdown of intergenerational communication has become the norm: the National Residents' Psychology and Aging Society Survey shows that more than 60% of adult children have less than one in-depth communication with their parents every week, and nearly 40% of the elderly say "they want to talk but can't find anyone", among whom the elderly living alone have less than 15 minutes of effective interpersonal interaction per day, far below the standard of more than half an hour required to maintain mental health. Long-term emotional deprivation makes the elderly more likely to seek virtual emotional compensation on the Internet.
Even if children are around, high-quality companionship is non-existent—when children accompany the elderly, their attention is often occupied by mobile phones; when the elderly want to chatter about daily trivialities, what they get in return is mostly perfunctory; a few more words of concern will be considered "nagging".
It is precisely this emotional absence that makes children habitually accuse rather than reflect when facing the elderly's addiction to AI videos. Little do they know that it is this dismissive attitude that gradually pushes the elderly into the "gentle circle" of AI emotional videos.
These low-cost, high-emotional-value AI emotional videos are essentially "spiritual opium". Liu Yanan, a professor at Communication University of China, points out that long-term reliance on AI emotional companionship will make the elderly fall into an "emotional cocoon room", reduce their tolerance for real social interaction, be unwilling to face complex interpersonal relationships, and even shift their emotional sustenance to virtual objects, weakening family relationships and forming the risk of emotional substitution.
To break this "emotional cocoon room", it is not by verbal preaching, but by using real companionship to refill the emotional vacancy of the elderly in reality. Real protection is never just a few words of "don't be deceived", but real daily companionship.
03. Algorithm Boost: How Platforms Become the Driving Force of Hunting
The emotional vacuum is only the "target", and what makes the sickle fall precisely is the continuous feeding of platform algorithms. If low-cost mass production is making bullets and the emotional gap is aiming at the target, then the algorithm is the hand that pulls the trigger.
The algorithm itself cannot identify "AI Overbearing President", it only identifies data. Liu Chang, an algorithm engineer from Douyin, publicly introduced at the 2025 Douyin "Security and Trust Center Open Day" that the algorithm predicts user behavior through neural networks, calculates the total value users get from watching each piece of content, and pushes the top-ranked content to users. The longer users stay on AI emotional videos and the more frequently they interact, the higher the system judges their traffic value to be, increases the recommendation weight, and continuously distributes them to more similar users. The algorithm cannot truly understand what high-quality content is, and it only judges based on likes, reposts, and completion rates.
With the content features of low information threshold and strong emotional hooks, the data performance of false emotional videos generated in batches by AI is significantly higher than the average value of ordinary content on the platform. According to Xinhua News Agency, an AI "overbearing president" account with 156,000 fans has a 5-second birthday greeting video that received 71,000 likes and 13,000 comments. Estimated based on the number of fans, its like rate is as high as more than 45%. Compared with Douyin's traffic distribution rules, the like rate threshold for new videos to enter the intermediate traffic pool is only 3%—which means that the interaction data of such videos far exceeds the distribution threshold of high-quality content on the platform, and they continuously get algorithm weighting.
The picture shows the comment section of the AI Overbearing President
Every stay, every like, and every sincere message from elderly users becomes an interest signal transmitted to the algorithm, prompting the platform to continuously and directionally feed similar content, and finally form a solid information cocoon room. The more you watch, the more it is pushed; the more it is pushed, the more you watch. The self-reinforcing traffic closed loop of the algorithm is thus formed, and the public domain customer acquisition cost of the gray industry chain is compressed to an extremely low level.
However, the algorithm-driven traffic carnival is only one part of this "hunting", and the absence of platform back-end risk prevention and control further allows the chaos to spread. On September 1, 2025, the *Measures for the Identification of AI-Generated Synthetic Content* jointly issued by four departments including the Cyberspace Administration of China came into effect. The regulations clearly require that all AI-generated content must complete explicit and implicit double mandatory identification. However, the full-platform actual measurement results of *Southern Metropolis Daily* show that most platforms have problems such as delayed automatic labeling and hidden risk prompts. Even if the label "Generated by AI" is marked below the video, the elderly fans in the comment section still believe in it firmly.
The AI content identification that should have built a solid risk defense line is in name only, and the loopholes in account banning governance make it difficult to eradicate the chaos. Data from Kuaishou's 2025 Community Governance Report shows that the platform has disposed of 89,000 pieces of illegal content and more than 2,900 illegal accounts.
However, according to an undercover investigation by Jimu News, the banned illegal entities have not completely exited. After an account is banned, operators quickly replace it with a new account to continue production. Practitioners revealed that "the maximum lifespan of a single account is 5 days". The speed of platform account banning is far slower than the speed of new accounts being launched, and governance has long been in a passive predicament.
Behind this governance predicament is the deeper contradictory position of the platform. On the one hand, the platform has the obligation to govern AI-generated content—removing illegal accounts, cracking down on false information, and labeling AI content. On the other hand, the traffic and commercial returns brought by AI content are assets that the platform cannot give up.
According to Kuaishou's 2025 Photosynthetic Creator Conference, compared with half a year ago, the playback volume of AI content on the platform in July 2025 increased by 321%, and Lingling AI's monthly paid revenue has exceeded 100 million yuan since April 2025. Gai Kun, Senior Vice President of Kuaishou, said bluntly, "AI has become a productivity lever for driving traffic and commercial monetization."
The platform is both the rule-maker and the beneficiary of traffic dividends. This kind of interest bundling is destined to make governance never keep up with exploitation.
From "fake Jin Dong" to "AI Overbearing President", emotional exploitation targeting the elderly has never stopped, and only the means have changed: AI makes video mass production close to zero cost, algorithms make emotional needs accurately captured, family absence provides a steady stream of "prey", and platforms are always swinging between traffic and governance.
Technology has lowered the threshold for doing evil, and business has amplified the efficiency of exploitation. Only those elderly people who can't even understand "generated by AI" are never protected.
The starting point to solve the problem is never stricter regulations, but to make the cost of violations higher than the benefits, and make the absent companionship be seen again.
This