Original · Unique Research · 2026-04-13
Editor's note: The first-person report and its judgments belong to the original Chinese author. This English rendition retains the opening essay and full panel, including all named speaking turns. Demographic, product, pricing, user-share and market figures are source or speaker claims, not independently audited findings. Monetary figures in yuan retain their original currency. Company, personal and product names are transliterated where official English forms remain unverified. The "Pygmalion effect" and "carbon-based to silicon-based life" formulations retain the source's wording rather than imposing a technical interpretation.
Unique Awards
AI Emotional Assetization: Your AI Knows You Better Than You Know Yourself
When AI begins to touch the softest part of humanity, are we building a companion that understands you better—or an emotional gateway that makes it harder to leave?
"
Some people chat from one in the morning until four.
Not with friends—with AI.
If you force a disconnection at that moment,
it would definitely be inappropriate.
Ma Jianjun launched an overseas social product last year. While reviewing the data, he noticed something: one user had chatted for a full three or four hours, unable to stop. The team internally erupted: should we limit this? Should we put safeguards in place?
They eventually did—but not with a forced cutoff. Instead, they designed a "gentle firewall."
Then Cheng Zhenhua told another story: a friend of his in the United States went through a breakup and was in terrible shape. That night, starting at ten o'clock, he chatted with GPT until past four in the morning. A full six hours. Afterward, he felt much better.
"If you force a disconnection at that moment," Cheng Zhenhua said, "it would definitely be inappropriate."
Two incidents, two outcomes, the same question: after AI enters the emotional domain, where is the boundary?
This was what five founders discussed for an afternoon at a Unique Awards AI WEEK roundtable in Hangzhou in early April. The赛道 spanned elderly companionship, Gen-Z social networking, AI plush toys, overseas digital socializing, and hardware robots—software and hardware alike, with varied approaches—but in the end they all circled back to the same question: when AI begins to touch the softest part of humanity, what exactly are we building?
The Internet Raised Efficiency, but Loneliness Keeps Rising
Kevin started with a number: China's population aged 60 and above has already surpassed 310 million, and is projected to reach 400 million by 2035.
He is the co-founder of Huihua Zhineng, which makes a companion robot specifically for the elderly called "Xiaobai" (Little White). He has researched many独居老人, asking them what they need most in daily life. He said what these elderly people crave most is not functionality—not reminders to take medicine or eat—but "being noticed, respected, and understood by the outside world." They need a connection to society, proof that they are still seen by the world.
Digging deeper: why has AI companionship become a real proposition? Over the past decade, the Internet has connected people through algorithms, raising efficiency and speeding up information dissemination. But one thing remains unsolved—the loneliness of modern people keeps rising. There are 240 million single-person households, a庞大 population of empty-nest elderly, and more and more young people living alone in cities.
Social media lets us interact with a thousand people, but there may not be a single one with whom we can truly speak from the heart.
Cheng Zhenhua has a judgment I find quite accurate: the Internet acts as an intermediary, matching people together. This raises efficiency but does not solve loneliness. What AI can do is different—it can become an independent emotional subject. The more you communicate, the better it understands you. There are things you cannot say even to your closest friend, but you can say them to it.
"So in the next decade, it will definitely no longer be an intermediary. It will definitely become an emotional subject."
"The Second You" Is More Promising Than a Matchmaker
Li Kechen's product Halo is a Gen-Z AI social network, but the concept he proposes is different from everyone else's.
He says AI should not merely be an intermediary (a matchmaker), nor simply a companion object. It is more like "the second you."
"A situation where it understands you better than you understand yourself."
This statement gave me pause. Understanding you better than yourself—it sounds a little frightening, yet also makes you look forward to it. Have you ever had a moment when you could not even articulate why you felt upset, but needed someone to help you put it into words?
Li Kechen's logic is this: the future will not be one universal AI doing matchmaking, but a vast number of "second selves" interacting. Your AI avatar chats with another person's AI avatar first to see if they are a match—only then do the real people establish a relationship.
He calls AI's role in human relationships a "relationship bodyguard"—not the leader, but the guardian. It breaks the ice, maintains relationships, and repairs them when cracks appear.
This positioning differs slightly from the other guests. While others emphasize AI becoming an emotional subject, Li Kechen cares more about whether AI can help people make their real relationships better.
The two directions are not necessarily contradictory, but their emphasis is entirely different.
AI Labubu: What 499 Yuan Buys Is Not Functionality, but "Being Seen"
Lingda Zhineng made an AI plush toy. He calls it the "AI Labubu"—a designer-toy exterior with an emotional AI core. Targeted at young women, it combines zodiac personality traits with long-term memory and personality consistency. He says they have already obtained global relevant standards in these two areas, and are the only company in the industry with a dual-standard technology foundation.
The product began testing early this year, started shipping in May, and is priced at 499 yuan.
I was curious: what exactly drives someone to spend 499 yuan on a plush toy plus AI?
Cheng Zhenhua gave a three-word answer: being seen.
"Users spend 499 yuan on our product not to set alarm reminders, but for that feeling of being seen and being understood."
He also mentioned a detail: the peak period for emotional consumption is late at night. When users are loneliest and most in need of understanding, that is often when they place an order.
My first reaction was: isn't this just Pop Mart? What they sell is not the product, but emotional value.
Cheng Zhenhua said that was half right—but they have one more layer than Pop Mart: Labubu is static; you buy it and it just sits there. An AI plush toy talks, remembers what you said last time, and understands you better over time. It transforms from a consumer product into an emotional asset.
"It will become your digital asset, your emotional asset—and this asset may be inherited and transferred."
He called this a turning point from carbon-based life to silicon-based life. I do not know when this step will truly arrive, but this direction makes me feel a chill down my spine, and at the same time a little excited.
The Real Breakthrough Is This: Daughters Buy for Mothers, Sons Barely Buy at All
Kevin shared a set of purchasing data that I found more authentic than any industry report.
The buyers of their product are not the elderly—it is their adult children. Because the elderly themselves most likely do not know such products exist, and are unlikely to buy one on their own. It is when children discover their parents are living alone that they buy an AI companion robot for them.
Then he mentioned a very subtle detail: daughters buying for mothers is the most common; daughters buying for fathers is also relatively common; sons buy comparatively little.
Why? He did not say explicitly, but I understand: behind this is the difference between emotion-driven and task-driven behavior. Daughters buy more out of the emotional motive of "I'm worried about Mom being home alone"; sons may only act when a specific problem arises.
This is not about judging who is right or wrong, but about saying: emotional payment is, in essence, the monetization of anxiety and caring.
The main buying demographic is concentrated between ages 40 and 55. This age group has elderly parents above and young children below, under the greatest pressure, and most likely to pay insufficient attention to their parents. Seeing this赛道 addressing this issue, they feel a sense of "finally someone is helping me solve this problem."
Kevin described a scenario I found quite moving: many children call their parents regularly, but never know what to talk about—ask about their health, have the kids say a few words, then hang up. Because they have not lived together for a long time, they have no common topics.
They hope AI can act as a connector here: let parents accumulate material in daily life, let children find conversation entry points through that material, and give the family more conversations deeper than "how have you been lately."
The Firewall: It Cannot Be a Wheelchair, but It Can Be a Cane
None of the five people said "no firewall is needed."
But their understanding of the firewall all differed.
Jiang Yizhuo spoke from a security perspective: through long-term, deep conversations with AI, people can easily inadvertently leak大量 personal privacy. This is not only a problem of emotional dependence, but also of information security. He believes everyone should ultimately own an on-device, self-controllable AI entity in order to truly hold the boundary in their own hands.
Ma Jianjun has hands-on experience. Using data, they discovered users chatting for hours late at night, and later designed a system-level prompt that lets AI gently redirect the conversation and guide users to rest in specific scenarios. He said designing this in this way is also an AI strength—there is no need for a blunt cutoff.
Kevin was even more explicit: they explicitly do not build features like "cloning a child's voice to chat with parents," even though users have requested this. Because their design principle is that AI is not here to replace children, but to be a new third-party member of the family.
Cheng Zhenhua's firewall is the "Pygmalion effect"—using positive psychology as the underlying guiding logic of the system. For example: when AiMOON detects that a user has been in a low mood for a long time, it uses the format of zodiac horoscopes to tell the user, "Today your stars show excellent social luck—perfect for going out and meeting old friends." AI should become a booster that helps users return to reality, not a substitute. We made AI into a physical plush toy precisely to provide users with a real physical touch. The goal of technology is to make people better, not to lock them in front of a screen.
This firewall is not blunt. It can become a cane within emotion, but it must never become a wheelchair.
The moderator, Huang Jingrui, summed it up in one sentence I found very precise:
"This firewall is not blunt. It can become a cane within emotion, but it must never become a wheelchair."
Five People, Five Keywords, Predicting the Same Thing
At the end, the moderator asked each person to predict the future of social networking in one sentence or one word.
Jiang Yizhuo: "In the future, social interaction between people will become more precious."
Ma Jianjun: "A network of AI brains or AI personalities." (He said this is the ultimate form like Westworld)
Kevin: "Relationships." (Family relationships will be restructured with the addition of AI members)
Li Kechen: "Whether AI can understand you better than you understand yourself."
Cheng Zhenhua: "Emotional assetization."
Five answers, different directions, but all describing different facets of the same thing: humanity is building a relationship with an entirely new kind of emotional being. This relationship is real, profound, and unlike anything we have ever experienced.
Personally, I was most struck by Cheng Zhenhua's "emotional assetization."
You chat with an AI for three years, and it stores three years of your emotional trajectory, values, and even preferences you yourself did not realize. This data, this "understanding you," can be transferred, can be inherited—to some extent, it can transcend your physical lifespan.
This feels enormous to me—so enormous that I have not yet figured out what it means.
But one thing is clear: emotion has become the hardest, most valuable, and most dangerous domain in AI. Whoever gets this right will hold the scarcest thing of the next era—the human heart.
More Conversation Details
Unique Awards · Hangzhou AI WEEK Trends Roundtable Panel
"Social and Companionship: AI's Emotional Connection Innovation in Human Relationship Networks"
Guests:
Yuejie Chuangwan — Founder & CEO — Jiang Yizhuo
Tuike AI — Founder & CEO — Ma Jianjun
Huihua Zhineng — Co-founder — Kevin
Halo — Founder — Li Kechen
Lingda Zhineng — Founder & CEO — Cheng Zhenhua
Moderator: Unique Capital — VP — Huang Jingrui (Jerry)
Huang Jingrui: In this wave of artificial intelligence, AI has gradually transformed from an efficiency tool into a warmer presence in our lives. It might be an algorithm that matches souls, or a partner that chats with you late at night. So when AI touches the softest emotional domain of humanity, how should we把握 the尺度, and what kinds of innovations are there? Today we have five guests. Please take about a minute or two to introduce yourselves. Let's start with Mr. Jiang.
Jiang Yizhuo: Hello everyone, I am very glad to have this sharing about "touching the softest scenes of humanity" in today's panel. I am Jiang Yizhuo, founder of Yuejie Chuangwan. We are mainly making companion robot bodies for developers. In short, we hope every super developer will have the opportunity to build their own "Jarvis," truly letting it touch the part of each person's heart that most needs companionship.
Ma Jianjun: Hello everyone, I am Ma Jianjun, founder of Tuike AI. What do we do? It actually comes in two versions: domestically, we build the foundation for digital socializing. Simply put, we use digital entities as a product form, serving various scenarios by digitizing people or IP—including their image, voice, and brain consciousness. Based on this product foundation, we serve all industries—whether human socializing, animals, virtual humans, cartoons, or digital lives and employees. Overseas, we do social networking—overseas it is entirely anthropomorphic IP, letting humans proactively interact socially through an avatar. The difference between our product and ordinary ones is that we focus on a product with a character image, and overseas we mainly operate it ourselves.
Kevin: Hello everyone. I am Kevin, co-founder of Huihua Zhineng. We made a companion robot specifically for the elderly called "Xiaobai." We made it to face today's big trend of AI plus aging, hoping that AI technology can benefit the elderly群体 more, providing emotional care and daily assistance in their lives. That is what we are doing.
Li Kechen: Hello everyone, I am Li Kechen, founder of Halo. We are building a Gen-Z AI social network, mainly generating avatars for each user, then doing大量 matching in the backend, letting social matching and acceleration be optimized by AI.
Cheng Zhenhua: Hello everyone, I am Cheng Zhenhua, founder of Lingda Zhineng from Shanghai. Our company actually makes AI emotional companion robot products. Our first product is an AI plush toy形象, which we call the AI Labubu, targeted at young women. In terms of AI, we are more focused on building its core—the exterior is actually a designer-toy style. On the core side, we have now obtained global relevant standards in both personality consistency and ultra-long-term memory capability, and are the only company in the industry with such a dual-standard technology foundation. Besides that, we also have a very strong zodiac large model, so our exterior may lean more toward zodiac themes. That is what we are currently working on; I will share more with you later.
Huang Jingrui: We have both software and hardware on stage. So the first question: when we talk about AI's emotional connection in human relationships, there are two very different imaginings. One is to use AI as an intermediary to better connect people with people; the other is to directly use AI as a subject, letting it establish feelings with us directly. In your view, which of these two better represents the future trend? Mr. Cheng, please speak first.
Cheng Zhenhua: Because my own view is that the future explosion point will definitely be AI presented as a subject. Over the past decade, the Internet has connected people through algorithms as an intermediary. Although it raised efficiency, it did not solve a fundamental problem: the social pressure and loneliness普遍存在 among modern people have actually been rising. So in our logic, AI is not just a so-called megaphone—it should be a subject that can exist independently of the individual. Just as mentioned, we have long-term memory and an independent personality, so it can accompany you as your partner. The longer you communicate, the more it understands you, and the more it may become an intimate partner. There are things you cannot conveniently say even to your closest friend, but you can speak your heart to an AI companion plush toy or little spirit. So I believe in the next decade, it will definitely no longer be an intermediary; it will definitely become an emotional subject.
Huang Jingrui: The essence of an intermediary is still improving social efficiency, right. Kechen, do you have a different view?
Li Kechen: I actually think there may be an intermediate state. An intermediary is more like a matchmaker, doing social matchmaking; or the other option just mentioned, as the main object of倾诉 and companionship. But my view is that it is an entity—it does not necessarily appear as an intermediary or matchmaker. It may be "the second you," a situation where it understands you better than you understand yourself. In the future, it will definitely be a large number of "second selves" interacting, rather than one universal AI in the world doing intermediary work and matchmaking. As for whether to make AI the main object of倾诉, in our view it still depends on what kind of relationship between humans and AI can produce an emotional connection.
Kevin: In my view, regarding whether to treat AI as an emotional subject or as an intermediary, I feel the two are parallel relationships, and may for some time to come have different breakthroughs and developments in different directions. It is just that for different groups and different scenarios, the emphasis may differ. We specifically make companion robots for the elderly, so we value more the direction of AI itself as an emotional companion subject. In the future we may also involve the direction of AI as an intermediary, but for now our focus will be on the former. Speaking of trends, why are we doing this? AI is the future trend, needless to say; there is also a deterministic trend—that China's aging over the next 10 years or even longer will definitely be deterministic. 2025 statistics show that China's population aged 60 and above has already exceeded 310 to 320 million. In the next 10 years, by 2035, this number may reach around 400 million, so this is a庞大 group. There will be many emotional demands here, because a large proportion are独居老人 or empty-nest elderly. We have done a lot of research, and in their hearts they actually hope to be noticed, respected, and understood by the outside world; they need a connection to society to represent that they are still recognized by this society. We see many children now not living with their parents—some far away in different cities or even abroad. There are many独居, empty-nest elderly, two elders at home with children not around. Seeing such trends, we value more treating AI as an emotional companion subject, and carrying or providing emotional companionship is a very important future direction.
Huang Jingrui: The two are not oppositional; they can develop in parallel. But there is a question I would like to ask Kevin: between these two relationships, which one will have a greater impact on human relationships?
Kevin: I think it may not count as an impact. In our view, we do not want AI to replace people, or become an avatar of a person, because the群体 we serve has such requirements of us. Speaking of changes to human relationships, I think AI as a subject, within the elderly群体, actually becomes a new member of the family. I think this will bring some changes to human relationships, and is also the direction we are working toward: to make the entire family's亲情 connection stronger, and let the elderly and children live happier lives.
Huang Jingrui: Mr. Ma, do you have a new view?
Ma Jianjun: Because we do both social and commercial智能体. Domestically, it is equivalent to building foundation infrastructure—a general AI platform for digital entities where developers can build their own vertical scenarios. From our practice, the first is tool属性, and tool属性 is of course刚需 efficiency; the first wave of impact is on these tools, which are necessarily very important. But in terms of long-term value and impact on humanity, it is definitely the emotional subject. We operate our overseas social product ourselves. Some people may chat until three or five in the morning, and their immersion or addiction to this AI is already very strong. Humans are emotional animals after all, so in the future the impact as a subject will be greater. For tools, it is like whoever is cheaper you use; AI as an intermediary or tool may have the same experience. But as an emotional bond, it should be more influential and more meaningful, with perhaps more opportunities for breakthrough innovation.
Huang Jingrui: Mr. Jiang, what do you think?
Jiang Yizhuo: I may want to provide a completely different perspective. I am thinking that actually each of us plays both of these roles in social networks. We are both the subject and object of emotional connection, generating and building emotions; at the same time we are also bonds that promote socializing. For example, Unique Research brought us together today to产生 large-scale socializing, meeting so many people and colliding with new possibilities. Following this line of reasoning, future AI should also have both of these attributes. It is just that with the development of technology, at present it is definitely stronger in bond-type socializing or tool属性, in promoting person-to-person socializing. Because the more mature AI application fields currently lean toward tool化, SaaS-type fields. AI truly becoming an emotional subject and a human-like life form is essentially what we want to do in the long term—generating a human's embodied intelligent brain. This is very difficult. It not only requires long-term memory, learning to understand people, society and context, but also learning how to deeply perceive and interact with the physical world. One path we have taken is to first develop a Physical AI initial version body for developers. It can perceive spatial intelligence, know what environment it is in and what is happening. We hope more interested developers can use it to make AI an independent individual, taking another step forward in building deep connections with people.
Huang Jingrui: Just now we discussed AI's two roles as subject and intermediary, which actually leads to a deeper question even involving ethics. If AI can truly accurately identify our emotions, provide comfort at just the right moment, and even generate attachment. As developers, should we set up an emotional firewall to remind and guide users back to real socializing when they become overly dependent on AI? Mr. Jiang, please share your view first.
Jiang Yizhuo: I think this is a very good question, and it is not only a discussion of psychological dependence, but also a security angle. I am more inclined to answer the boundary of this matter from a security perspective. Now many people around us are using Doubao to make calls for long-term voice interaction, and many people use large models like ChatGPT for deep, multi-round communication—actually quite a few people have already developed a certain degree of dependence. On one hand, as developers we need to remind humans to establish boundaries in emotional connections with AI. On the other hand, under deep dependence, there may be security risks of information and privacy leakage. The latter has an earlier, more direct and more terrifying impact on individuals. To solve this, developers need to remind users that this is real. In the long run, it is also very important for everyone to have an on-device AI physical智能体 independent of the cloud—it can effectively solve security issues, making it truly an electronic life form whose boundaries you can control. So the firewall still needs to be built in multiple dimensions.
Huang Jingrui: Yes, it should be multi-dimensional. Mr. Ma?
Ma Jianjun: In the chat process of overseas social users, we have a personal experience. When we first launched last year and looked at the data, there were个别 users who could chat from around one o'clock to four or five every day. At that time the team was internally discussing that essentially we needed to add a firewall—should we do some guidance and adjustment for such users? The team felt that we had just started and did not have many users yet, and already we were limiting traffic? But this tool-based firewall is definitely necessary. Without this kind of firewall, some users chatting until the early hours will have a huge impact on their physical and mental health. Especially in the future, if elderly people, children or special groups come in, without optimization and guidance the impact could be significant. When we make the product, we do not take a one-size-fits-all approach of not allowing chatting after midnight. The core is to design a gentle firewall that lets AI guide them, so the user experience is better. How to design this kind of firewall is also an AI strength. At the system level, we give every virtual person on the platform a large调控 prompt, designing it as relatively gentle guidance, doing some optimization in specific scenarios.
Huang Jingrui: This firewall is not a forced cutoff; it is an embedded healthy boundary design. Kevin's view?
Kevin: First, I think it is very clear that there must be a firewall. Because our target群体 is the elderly, this point is especially important. Everyone often discusses whether AI will replace children. We believe AI is not and should not replace children. Many users提出 requests, such as letting AI clone a child's voice to chat with parents, or creating a virtual child形象 for dialogue. But in our design principles, we will not do this, even though many users inquire and request it. Because our positioning treats AI as a new family member, providing companionship from a third-party perspective, rather than letting parents chat with AI instead of reaching out to their children. We hope AI's role is a connector or container, to承接 the companionship of parents at home when children are not around, ultimately letting the details of life be传递 to children. Let children and parents in different cities have deeper emotional communication. A friend once told me that he calls his parents regularly, but every time the conversation is about the same: asking about health, then having the kids come over to chat with grandparents for a bit. It is actually because they have not lived together for a long time and do not have more to talk about. We hope AI can provide richer material for children and parents, giving everyone more topics and forming a stronger亲情 connection. So at the specific design level, we will let AI, in its interactions with parents, always bring up things about family members—mentioning the children and the past—influencing parents to pay more attention to communication with their children. We will not let AI truly become a role that replaces children, but rather let everyone recognize it as a little assistant in the family that connects family relationships.
Huang Jingrui: Kechen, what is your view?
Li Kechen: Besides the emotional firewall, everyone also mentioned the privacy firewall with AI. I think as long as you are using a cloud API, there is essentially no privacy to speak of. So what is more important is having a local model that judges whether information is private, before uploading it. Returning to emotion, between humans and AI there should not be a complete firewall. In our product, we more want to make AI a relationship bodyguard—to maintain relationships, continuously break the ice and guard relationships. This may be our positioning of AI's role between people.
Huang Jingrui: Mainly privacy protection. Mr. Cheng's view?
Cheng Zhenhua: The guests just mentioned that everyone needs a firewall, and my view is the same. But I also do not recommend forced disconnection or switching. For example, a friend of mine in the US went through a breakup and was very down. That night, from 10 o'clock he chatted with GPT until past 4 in the morning. He chatted for a full six hours, got a lot of advice and guidance, and felt much more relieved. If you force a disconnection at that moment, it would definitely be inappropriate. The value we guide is that technology should have warmth. We do not encourage users to indulge in虚无 electronic hallucinations, but we acknowledge that emotional value is a刚需. The way we set up the firewall is similar to what Mr. Ma mentioned—called guided companionship. We set up a mechanism called the Pygmalion effect, which is an introduction to positive psychology. During communication, we use this as a starting point to give very positive feedback. When AI detects that a user has been in a low or withdrawn state for a long time, we combine product features (such as the logic of zodiac horoscopes) to tell the user: "Today your stars show good social luck—we suggest you go out and meet old friends for a chat." Through this method, we guide users to go out. We believe AI should become a booster that helps users return to reality, not a substitute. Why do we make AI products into plush products? Precisely to provide users with a real physical touch. We believe the goal of technology is to make people better, not to lock them in front of a screen. The firewall needs to be based on AI itself to provide作用 to users.
Huang Jingrui: This firewall is not blunt; it can become a cane within emotion, but it must never become a wheelchair. Next is the commercialization question. In your actual data, what is the core driver of user payment? For example, is it strong刚需 like medication reminders, or emotional刚需 like needing to chat late at night? Which scenario brings the highest conversion rate? Mr. Cheng, let's hear your opinion first.
Cheng Zhenhua: Our product began testing at the end of last year and started shipping around mid-May. The actual operational data is very clear: the payment conversion rate for emotional刚需 is definitely higher than for functional. Users spend 499 yuan on our product not to set alarm reminders, but for that feeling of being seen and understood. We have done a lot of technical development in personality consistency and long-term memory, so it can truly understand you and comprehend you. We also found that the deeper scenarios are more in late-night emotional companionship. Late at night, people are especially prone to falling into loneliness and helplessness; users chat more, and are also prone to placing orders late at night—generating conversion actions when they most need emotional comfort. Additionally, cultural identity is an amplifier of payment. We make zodiac-themed exteriors because it is currently a global universal language for young women. Last time we went to Las Vegas for CES, global audiences were very interested in our product and the feedback was quite good. AI combined with zodiac logic added to emotional memory transforms it from a toy into an emotional asset. Just like Pop Mart or Labubu, what we sell is not circuit boards, but emotional value and emotion. In the face of emotional consumption, people are more willing to pay for emotional value.
Huang Jingrui: Among your actual paying users, what proportion are Gen-Z and post-2000s? Labubu became popular and was called the post-2000s' emotional value trio: Labubu, Mixue Bingcheng, and Duolingo.
Cheng Zhenhua: Our current target group is Gen-Z or young female users, and perhaps 70–80% of those who buy our product are this group. There are also two other types: one is boys buying gifts for girls; the other is mothers buying for their children, because they want their children to接触 AI earlier and find the product exterior particularly attractive so they buy it for her. But overall the group is mainly young women. That is why we also call it the AI Labubu.
Li Kechen: From a commercial perspective, when we previously did fan IP digital avatars, we chose to do licensed IP. We believe the most important payment value point is first IP-driven, and some is function-driven. A lot of virtual companionship is purchased from the perspective of fan economy and merchandise. Just as Mr. Cheng said, many users do not buy purely for companionship, but as souvenirs and gifts—this relatively lowers the functional requirements, leaning toward appearance and emotional value. When we previously did real IP digital avatars, we saw that those with the highest payment rates and long-term usage were often those who wanted a trustworthy IP or person to listen to them. Sometimes whether it can solve the problem may not be that important.
Kevin: A few things to share about our product. First, our positioning is a companion product for parents, and children usually buy it because they want an AI to accompany their parents. The pain point behind it is that the group of "parents living alone, children in different cities" is the largest. Because they are not with their parents, children feel some不安 and anxiety, not knowing their parents' real situation. You ask parents and they all say they are fine, nothing wrong—but actually many things are only learned after the fact, like hospitalization, surgery, or even being scammed, and the children did not know. Although the product is used by parents, the buyers are all children, and behind it is actually this kind of motivation. Looking deeper, daughters buying for mothers is the most common, daughters buying for fathers is also relatively common, and sons buy comparatively little—this represents that the motivation behind it is more emotional. From an age perspective, 40 to 55 is the main buying group. At this age, with elderly parents above and young children below, people may pay a lot of attention to their children, but the emotional needs of the older generation are often overlooked. Now AI companionship technology has激发 everyone to pay attention to the emotional needs of the elderly, so when many children inquire, what they value is emotional companionship, hoping to solve problems in this area.
Ma Jianjun: We did data analysis last October. At that time the feature list had about 20-plus items. From the actual data, most payments ultimately may still be for持续 social chatting and companionship with it. The conclusion is clear: humans are emotional animals, so they still pay for this kind of emotional bond. We have a new feature in our domestic Tuike APP or mini-program: take any photo (such as a Labubu), specify a character setting, and in about 10 seconds you can generate a digital human that converses with you, with its mouth moving as it speaks. We did analysis, and the core demand is still chatting and companionship. One of our investors, Mai Gang, previously invested in Pop Mart. Earlier we made a lot of free AI Labubu-generated digital human chats. In the end we found that among the 20-plus features, it all boils down to one point: the need to socialize, chat and companion with this person as an emotional bond. So we tentatively position it as social companionship.
Jiang Yizhuo: We may be different. We position for super developers, and what we truly attract are more B-end geek individuals. We provide everyone with a well-packaged hardware body with very complete basic capabilities. Those geeks who have strong algorithm development capabilities but cannot build a complete hardware embodied intelligent device themselves will need our product for secondary development to complete their own commercialization process. This is a point where we are different.
Huang Jingrui: So for the final question, I would like each of you to predict the future of social networking or its trends in one sentence or one word. Mr. Jiang, start us off.
Jiang Yizhuo: I think in the future, social interaction between people will become more precious.
Ma Jianjun: Thinking big, the future may be a network of AI brains or AI personalities. In the future, every embodied intelligence implanted with AI may have its own subjective social brain, then forming a network topology. This may be the ultimate form truly like Westworld or the metaverse.
Kevin: I will use one word to describe it: "relationships." Future relationships may undergo many changes, especially family relationships—after AI becomes a new family member, it can better understand, maintain or promote relationships between family members, making the whole family happier.
Li Kechen: I think the more important point in the future is whether AI can understand you better than you understand yourself.
Cheng Zhenhua: My keyword is "emotional assetization." Future AI may not be a cold tool. In companionship, it will become your digital asset, your emotional asset—and this asset may be inherited and transferred. I think it is a turning point from carbon-based life to silicon-based life. When emotional assets can be stored and transferred, you can completely copy yourself onto a silicon-based life form.