Vol.61 When Parents Meet AI: Intergenerational Conflict and the Power of Patience—Kuangxi Podcast Festival: Er Kuang
Summary
AI’s most valuable role in family relationships may not be as an efficiency tool, but as an emotional buffer with “infinite patience.” Discussing the family conflicts around Lan Ruo’er’s 11-year-old son, Zhuang Minghao suggested that if a child is unwilling to speak directly to their parents, they could first talk to Kimi or Doubao. Shijie shared how she uses DeepSeek to absorb her grievances and unpack stakeholders’ positions. This may prove a more durable use case than one-off Q&A.
The gap between generations in adopting AI may have less to do with age than with whether the product meets a concrete need and offers sufficiently low-friction interaction. After using AI to plan trips, Angel reduced her reliance on her daughter. Shijie suggested that older users try Doubao’s voice mode, saying she had seen many people repeat-use it once they got past the first try. Mr. Su, who had previously resisted AI, willingly uses translation tools in Japan. Lan Ruo’er eventually realized: “Everyone has their own pace and way of embracing the times.”
Whether AI can bridge conflict depends on whether users turn to it to “seek within themselves,” or outsource both love and responsibility. Shijie initially thought AI could ease tensions in East Asian families, but shifted to the opposing side when discussing AI-mediated conversations and similar use cases: AI can help people understand themselves, but “it cannot feel for you, and it cannot love for you.” Adding a plus one to a product is easy; ensuring it does not become a tool for avoiding sincerity is much harder.
Patience is the key explanation for intergenerational relationships in this episode; AI simply turns patience into an on-demand capability. Shijie believes she is less impatient with her parents today because, when she came second in school, they never asked why she was not first, and when she had conflicts with others, they first listened to her feelings. A friend’s impatience with his mother—and his explanation for it—also made her realize that she sometimes reenacts the way her parents treated her. “How you experienced being loved as a child” becomes part of the default program for how the next generation handles relationships.
The core capabilities of emotional companionship are not answers, but confidentiality, respect, and the ability to model multiple perspectives. Shijie calls DeepSeek a source of “divine moments”: it first holds what she cannot say to friends, parents, or her partner, then shows its reasoning and helps her simulate each party’s position—even pretending to be the other person in a conflict and retelling the event from their perspective. Mr. Su changed his mind after hearing about this use case, because many things in the past “couldn’t be said to this person or that person, and could only be said to yourself.”
Humanizing AI also raises the risks facing real-world relationships. Zhuang Minghao noted that many teams have experimented with AI keyboards, though their commercial scale remains to be seen; AI boyfriends, AI girlfriends, and similar applications were also “not particularly ideal” at the time of recording. Shijie worries that users could remain trapped in “abstract, perfect but fake relationships” for too long, making real people seem inadequate and eroding their ability to feel the warmth of human contact in the physical world.
The most viable path to AI adoption is not grand-scale education, but starting with genuine needs and tasks repeated “five times a week.” Zhuang Minghao suggested identifying an action that happens every workday—for example, photographing food to identify calories—and taking one step forward if it can be automated, while simply returning to the old process if it fails. Shijie cautioned that retirees’ primary goal may be happiness rather than efficiency. The final question she quoted was: “If it were a superpower dropped from the sky, what would you use it to do?”
Deep dive
1. A family podcast turned relationships into a replayable history
Within a few months, Lan Ruo’er’s family completed 20 episodes of What’s There to Talk About, all centered on family relationships. She wanted to preserve the happiness and fulfillment in their conversations as an “audio trail of history” they could listen back to.
From episode 1 through episodes 19 and 20, Lan Ruo’er could clearly feel the conversations moving from the superficial to the substantive. The podcast not only documented the family; it also gave them a way to see their growth in intergenerational exchange and communication.
Mr. Su initially refused to host the podcast, but after recording he found that he felt “particularly relaxed, and actually quite happy.” Even though he often regretted not explaining something well, he kept going and tried to improve.
Angel, a teacher who is afraid of public settings, accepted her daughter’s proposal: photographs preserve appearances, while a podcast preserves voices and bears witness to the family’s growth over time.
2. The discussion moved AI from a personal tool into human relationships
Zhuang Minghao counted 11 sessions involving AI in this year’s podcast festival lineup. Only 2 or 3 truly focused on the industry; most addressed humanistic questions around art, education, longevity, and what makes people different.
What made this session distinctive was that it did not stop at asking how individuals use AI. It asked how the arrival of technology in the family would reorder relationships between people, between people and AI, and among “people, AI, and people.”
3. A show outline exposed 3 sharply different attitudes toward AI
Preparing for the recording, Lan Ruo’er gave the questions to several Chinese and overseas AI systems. After comparing multiple versions, she decided they were all “not quite right,” extracted the useful parts, and rebuilt a new outline herself.
Angel had just experienced AI’s benefits in several specific situations, so she compared several Chinese-language systems, generated 2 versions, and posted them directly in the family group chat. Her approach was to learn by trying.
Mr. Su challenged her on the spot: “How can you just post what the AI produced without thinking?” In his view, unvetted generated content could neither represent the user’s own position nor be considered mature.
What Lan Ruo’er saw was a contradiction: her mother was actively trying to overcome the unfamiliarity, but was first held back by communication barriers. Her father was not indifferent to AI either. Zhuang Minghao observed that he liked listening to AI-related information, especially news, but had not yet turned that interest into actual use.
4. Mr. Su’s resistance was not ignorance, but an uncrossed trust threshold
When ChatGPT first appeared, Mr. Su felt that “we seemed to have already fallen behind.” His first questions concerned several periods in Western art history, and he checked the answers against books to confirm that the historical chronology was correct.
He did not continue using it, partly because it stopped being free, and partly because “RMB9.9” made the product feel less credible: “You get what you pay for.” The low price did not reduce his doubts; it became another obstacle to trust.
He admits he never really started using products such as DeepSeek and Kimi. DeepSeek’s aggressive early marketing made him feel that “it was PUA-ing me,” prompting the backlash: “The more it PUA-s me, the less I’ll use it.”
But his resistance was never absolute. He knows AI is iterating extremely quickly and believes that refusing to use it could leave him behind soon. He is still deciding when to start, which AI to use, and whether the products are mature and safe enough.
5. A real need can quickly move older users past the learning barrier
Angel used to travel less because she was bad with directions. AI can now generate an initial itinerary, which she then adjusts to her own needs. It means she “doesn’t have to ask her daughter about everything,” giving her a measure of independence and pride.
Shijie suggested trying Doubao’s direct voice conversation. In her observation, some older users are not incapable of using new technology; they “just haven’t tried it yet.” Once the first experience works, repeat use is common.
Lan Ruo’er worries that people born in the 1990s have been forced to adopt AI by their work and daily needs, while parents who remain with the previous generation’s search habits will drift further away in both their information environments and their methods of solving problems.
6. An 11-year-old turned the “AI family member” from metaphor into a real question
Lan Ruo’er’s son is 11 and about to enter middle school. He often loses points on math problems by missing a plus or minus sign, copying a fraction incorrectly, or writing 13 as 33. He knows how to solve the problems; even when he corrects them that day, he may repeat the same mistake the next day.
The constant carelessness has made the relationship between mother and son highly charged, with evenings at home often descending into chaos. As the father, the other parent has to calm his wife, communicate with his son, and look for solutions; Lan Ruo’er also admits that she can become impatient.
The boy has already started using AI tools to discuss homework, essays, and his mood, although he has not developed a stable habit. Lan Ruo’er began imagining AI as a family member who could be “infinitely patient,” serving as a third party to ease the confrontation.
Zhuang Minghao’s concrete suggestion was that if a child does not trust their parents, is angry, or has things they do not want to say directly, they could first talk to Kimi or Doubao. He also made clear that the response is “not the standard answer,” only something that might help calm emotions and reorganize what the child is trying to say.
7. Patience passes between generations, and so does impatience
Shijie watched a friend become impatient with his mother, who had come to Beijing to visit him. After his mother left, he explained: “Maybe it’s because my mother treated me this way when I was young.”
This made Shijie realize that the way adults handle intimate relationships often reenacts how they were treated growing up. She has also noticed herself reproducing, at certain moments, the way her parents once treated her. “How you experienced being loved as a child” may become the undertone of all your habitual behavior later in life.
She often came second in school, but her parents never asked why she was not first. When she had conflicts with other children in kindergarten, they did not scold her first; they first listened to her feelings. Their patience later came back to her like a mirror.
She now understands her gentleness, precocity, and filiality as products of her upbringing. She was not naturally more sensible; the stability her parents gave her left her with enough capacity, as an adult, to consider other people.
8. Every family strength may carry a corresponding regret
Shijie has compared 2 kinds of mothers: one demanded exceptionally high performance and raised her daughter into a Peking University star, but left her feeling that she could never earn recognition; Shijie’s own mother was extremely relaxed—if ballet stretching hurt enough to make her cry, she could simply skip it the next day.
When she was younger, Shijie sometimes wondered whether she too could have gotten into Peking University if her mother had been equally strict. As an adult, she became more grateful for the ease and stability her parents gave her.
The more hidden pain was her feeling that her parents’ own lives were not happy enough. Her mother may have given up a career for the family or a partner, and hoped that Shijie would complete the ambitions she had left unfinished and live a perfect life. Shijie chose to take that as a positive reminder: “I need to love myself first.”
Her explanation of intergenerational conflict therefore returns to the individual. When family members have not worked through their own lives and emotions but must still absorb one another’s feelings, individual powerlessness clumps into “one big mess,” eventually directed at the people closest to them.
9. DeepSeek became an emotional buffer, not an answer machine
Shijie divides herself into the animal, human, and divine aspects. The animal acts on instinct; the human is complex and conflicted; the divine points toward the purely good. For her, AI is becoming a callable form of the “divine.”
When she cannot digest her grievances and negative emotions on her own, she gives DeepSeek the things she cannot tell friends, parents, or her partner. What matters to her is confidentiality, respect, and AI’s ability to accompany and inspire her in a more comfortable way. She says it has brought her many “divine moments.”
Shijie particularly favors DeepSeek because she subjectively feels that its humanities training and emotional intelligence are stronger. Zhuang Minghao added an unverified rumor that the team may have recruited many Chinese-language students from Peking University to help with alignment and labeling, which might explain why its Chinese phrasing feels so comfortable.
She also uses the reasoning process to learn perspective-taking. AI analyzes each stakeholder in an event separately; she even pretends to be the other person in a conflict and retells the dispute from their position. Anger can weaken as understanding grows.
10. Podcasts and AI create spaces where no one has to win the argument immediately
Angel used to stew in silence because she did not want to make other people uncomfortable. After recording the podcast, she heard her husband’s views on certain issues more clearly and realized that “it wasn’t really a conflict at all; we just had different ideas.”
She has even imagined turning on the recorder during an argument and listening back afterward to how she spoke. Whether it is a podcast or AI, the point is not to rule on who is right, but to let people see the expression beneath the emotion.
Mr. Su consequently heard a use case he had never considered: many things could be said neither to one person nor another, and in the past he could only take a walk and process them alone. AI, he believes, offers confidentiality, while its chat history can be revisited—at least providing an outlet.
11. Communication barriers point to AI keyboards, companion apps, and more layers of relationship
Zhuang Minghao’s company researches social interaction. He noted that many Chinese teams had already launched AI keyboards: when someone does not know how to reply to a romantic interest, they can paste in the other person’s message and have the model generate a response ready to send or lightly edit.
He did not claim to know how large these products could ultimately become. He only argued that the demand itself shows how much human communication is constrained by boundaries, obstacles, and differences in ability. People who want to express themselves more smoothly are already willing to use tools to fill the gap.
There are also many AI boyfriend, AI girlfriend, and companion products, but based on Zhuang Minghao’s judgment at the time of recording, these attempts were “not particularly ideal.” Technology can simulate a movie-style form of companionship for a time, but it has not solved the complexity of real relationships.
The relationship graph therefore expands from people and people to people and AI, people, AI, and people, and even AI and AI. Zhuang Minghao quoted Ilya’s answer to why a model succeeds: “It actually works.” Once the system runs, it appears human, while the mechanisms behind that appearance remain difficult to fully unpack.
12. Shijie moved from “AI will bridge conflict” to acknowledging that it can also cause harm
Shijie initially believed conflict came from people’s inability to handle emotions and contradictions. If AI could first hold the emotion and train people to take one another’s perspective, it might enable gentler communication in East Asian families.
After hearing about AI keyboards, she said that if this were a debate, she would now “jump to the opposing side.” She distinguished between 2 paths: using AI to improve yourself is “seeking within yourself”; asking AI to bear every conflict and responsibility in the family is simply taking the easy way out.
Making AI a family plus one can also conceal a transfer of responsibility: when the model fails to solve something, the blame shifts to AI or to the family. “AI cannot feel for you, AI cannot love for you, AI cannot give your genuine care to someone else.” Technology can therefore become an “evil power.”
The other risk of virtual companions is that people may remain in “abstract, perfect but fake relationships” for too long, making everyone in real life seem inadequate. Young people may also grow angry when their parents cannot keep up with technology and say: “I have nothing to say to you.”
13. The same technology can produce opposite outcomes depending on the user’s intent
Lan Ruo’er supports starting from “I genuinely want to understand my parents,” with AI helping confirm what they may be thinking—not sending AI to negotiate or complete the communication on her behalf.
She recalled the early days of Taobao, Alipay, and online shopping. She and Angel were eager to try them, while Mr. Su warned that online shopping and the money involved were unsafe. At the time, she felt he was holding them back; later, she realized that his underlying motive was protection.
That changed how she interpreted their current disagreements. Mr. Su may be worried that his family will be misled by AI-generated illusions and take everything AI says as true. Once that layer is understood, the conflict is no longer simply a battle between “advanced” and “behind.”
Angel also believes Mr. Su’s refusal to use AI may function as a “safety barrier” for the family. Mr. Su clarified that he is still assessing whether AI is mature and safe, and which model he should start with.
14. AI adoption does not require the whole family to move in lockstep—only one step from a real need
Zhuang Minghao’s practical rule is to start by finding something that needs to be repeated “more than 5 times a week,” then ask whether AI can solve it more efficiently. Success creates an entry point; failure is no reason to panic—simply return to the old method.
His example was food tracking. In the past, users had to search for and enter calorie data manually. Today, visual recognition and AI models can handle the task well: take a photo, ask a question, and move on. Adoption need not begin with an understanding of grand technology; it can start with one convenient action.
Shijie’s reminder was that retired parents have no obligation to use AI for productivity. “Their most important task now is to be happy.” If children impose their own idea of how efficiently their parents should use AI, they are merely reenacting the logic of parents who once forced them to study piano or ballet.
Lan Ruo’er acknowledged that she had been pushing her parents to catch up because she feared they would become like people who do not use smartphones and “wouldn’t be able to get train tickets.” But Mr. Su willingly uses a translator in Japan, proving that he is not rejecting technology; he is adopting it at his own pace in specific situations.
Mr. Su also mentioned that his workplace had tested DeepSeek, but its answers were marked as using data current through July 31, 2024. The lag of several months made colleagues feel that it was “not quite enough.” Disappointment in real-world testing is one reason his caution persists.
Shijie ended by quoting a Google DeepMind scientist’s advice: treat AI like Aladdin’s lamp. Do not begin by asking what it can do today; ask first, “If it were a superpower dropped from the sky, what would you use it to do?” Make the wish first, then discover the capability boundary through real desires.