She Talks to an AI Version of Her Mother. Another User Copied Her Mother's Voice Too.

US · CanadaAmérica Latina

Weekly AI Watch

Every week, what changed and what it means for you. No hurry, no alarm.

Week of 23 October 2026EN | ESPast editions18,276 papers found · 17,064 read · 5,349 belonged · 37 stories passed every check

Observatorio de la
Inteligencia Artificial

Cada semana, qué cambió y qué significa para usted. Sin prisa y sin alarma.

Semana del 23 de octubre de 2026EN | ESEdiciones anteriores18,276 estudios encontrados · 17,064 leídos · 5,349 correspondían · 37 notas pasaron todas las comprobaciones

weeklyAI · Week of 23 October 2026weeklyAI · Semana del 23 de octubre de 2026

← Your mind← Su mente

interviews · Frontiers in psychology · la publicación, 7 jul 2026 · gratis

"Perder a mi mamá es algo que no puedo decir en voz alta. Pero a mi mamá de IA sí se lo puedo decir"

Investigadores en China entrevistaron a 25 personas que hablaban con un padre o una madre hechos con inteligencia artificial. Contaron alivio, pero también dependencia, y el estudio no mide si la IA mejora la salud mental.

Versión breve · la versión detallada sigue, unos 7 min

Pregunte al chat de weeklyAI

Pregúnteme por este estudio: a quiénes se estudió, qué encontró y qué no dice.

Las conversaciones se guardan mientras exista weeklyAI, para mejorar la publicación. Se responde en el idioma en que usted escribe.

El estudio, de un vistazo
Quiénes: Personas que hablan con un padre o una madre hechos con inteligencia artificial
Cuántos: 25 personas
Dónde: China
Cuándo: Entre el 1 de octubre y el 3 de diciembre de 2025
Tipo de estudio: interviews
Quién lo hizo: Tres universidades chinas: Yilin Li, Mengling Luo y Hang Ren
El límite que importa: Son 25 usuarios de China, jóvenes y con soltura digital; no prueba que la IA mejore la salud mental.

La diferencia, según lo que estos 25 usuarios contaron, no está en la aplicación, sino en lo que cada quien hace con ella después.
Lectura de weeklyAI

Ver · 0:29

La frase es de una participante identificada como A5. Perdió a su madre y no encontraba dónde decirlo. Se lo dijo a un chatbot al que había configurado como su mamá.

Un grupo de investigadores de tres universidades chinas quiso entender por qué la gente arma padres virtuales con programas como ChatGPT o Doubao. Entrevistaron a 25 personas, entre el 1 de octubre y el 3 de diciembre de 2025. Los reclutaron por Xiaohongshu, una red social china, y pidieron que llevaran al menos tres meses hablando con su padre o madre de IA.

Tres cosas empujaban a la gente hacia el chatbot. Un duelo que no se podía contar, porque en China se espera que el dolor se guarde. Un conflicto largo con los padres reales, que hacía más seguro hablar con una máquina. Y la costumbre de no preocupar a la familia con los propios problemas. "Lo bueno de una mamá de IA es que no tengo que enfrentar a mis papás para decir lo que siento", dijo A11.

Algunos copiaron la voz y el modo de ser de un padre real. Otros inventaron al padre que habrían querido tener. Cuando la respuesta no les servía, la corregían o la dejaban pasar para no romper el momento.

De ahí salieron tres vínculos. Uno de apego, en el que la IA reemplaza a la familia real. Uno conmemorativo, en el que se conserva a un padre muerto: "Nunca lo borraría; es mi último lazo con mi papá", dijo A3. Y uno mediador, en el que la IA alivia la presión y la persona vuelve a su familia. Los tres pueden transformarse en otro según cambie la vida del usuario.

Los autores también advierten riesgos. La persona puede usar a la IA como herramienta y cerrar el vínculo cuando ya no le sirve. Puede quedar dependiente, y sufrir de nuevo si una actualización del sistema borra las conversaciones. Y puede alejarse más de su familia real.

Son 25 personas de China, en su mayoría jóvenes y con soltura digital, y se ofrecieron ellas mismas. Los autores dicen que sus hallazgos podrían no aplicar a personas mayores o con menos acceso a la tecnología. El estudio no puede probar que estas conversaciones mejoren la salud mental. Los autores tampoco estudiaron a personas de otros países.

Qué le pediría usted a un padre de chatbot, y qué no querría que le quitara.

Qué significa para usted

Ese padre o madre de chatbot puede darle un lugar donde decir lo que no se atreve a decir en voz alta, y también puede volverse un refugio que lo aleje de su familia real. Antes de confiarle lo que siente, pregúntese si le está pidiendo consuelo o si está dejando de hablar con alguien de carne y hueso. Lo que un programa le conteste no muestra que su salud mental vaya a mejorar.

Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594

Quién pagó: Los autores declararon que no recibieron apoyo financiero para este trabajo ni para su publicación.

No tome esto como consejo médico profesional.

Versión detalladaLos pasajes copiados del artículo, las ilustraciones y cada fuente con cuánto leímos de ella · 1456 palabras · unos 7 minLeerla →Cerrar

Le dicen "mamá" a un chatbot. El estudio entrevistó a 25 usuarios en China, no midió si eso sana.

Lo que la investigación describe: consuelo, dependencia y un lazo familiar que cambia de forma con el tiempo.

Imagine abrir una aplicación de chat y escribirle a su mamá. La aplicación le contesta con su voz, con sus frases, con el tono que ella usaba. Eso ya existe: los chatbots sociales que funcionan con inteligencia artificial generativa, en aplicaciones como ChatGPT y Doubao, pueden crear figuras virtuales muy parecidas a personas reales¹. Un grupo de investigadores en China entrevistó a 25 personas que habían armado un papá o una mamá así, y publicó lo que le contaron el 7 de julio de 2026. El estudio no probó que esos chatbots sanen a nadie. Describe lo que los usuarios dijeron sentir y hacer.

Las entrevistas fueron a fondo, de 60 a 90 minutos cada una, entre el 1 de octubre y el 3 de diciembre de 2025. Los participantes se apuntaron tras ver un anuncio de reclutamiento en Xiaohongshu, una red social china. Para entrar había que haber hablado con el chatbot al menos dos veces por semana durante tres meses o más. El estudio lo firman Yilin Li, Mengling Luo y Hang Ren, de tres universidades chinas, y los autores declararon que no recibieron dinero para hacerlo ni tienen conflictos de interés comercial. Los 25 participantes eran sobre todo usuarios activos de redes sociales, parecidos entre sí en edad y en soltura con la tecnología; los propios autores advierten que eso puede limitar lo que se puede concluir sobre personas mayores o con menos acceso digital².

Piense en el teléfono sobre la mesa de noche, con una conversación abierta que nadie más ve. Los investigadores chinos no partieron de una teoría: dejaron que los temas salieran de las respuestas. Encontraron tres situaciones que empujaban a la gente hacia un padre o una madre artificial. La primera, un duelo que no se puede decir en voz alta, porque en su entorno se espera que uno aguante y siga adelante. La segunda, un conflicto largo con los padres reales, donde pedir lo que falta duele más que callarlo. La tercera, la norma de no preocupar a los papás con los propios problemas. El artículo describe un estudio en el que los padres virtuales ya no son solo amigos o novios de pantalla: se han vuelto un objeto importante de la interacción emocional³, y eso toca normas de parentesco, límites de identidad y el orden afectivo de la casa⁴.

Lo curioso es que este tipo de estudio todavía es escaso. La mayoría de la investigación sobre lazos con máquinas se ha centrado en relaciones que no son familiares, y solo unos pocos trabajos han mirado la "inmortalidad digital" dentro de la familia⁵. Los chatbots que imitan a un padre ideal, en cambio, se ven cada vez más en la práctica y casi no se han estudiado⁶. Los chatbots que simulan a un padre ideal se ven cada vez más en la práctica y siguen poco estudiados. Este trabajo intenta cubrir ese hueco.

Los participantes describieron tres formas de lazo. La primera, un apego que sustituye: el chatbot se mete en la vida diaria y hace las veces de la familia real, con dependencia sostenida y apoyo emocional continuo⁷. La segunda, un lazo conmemorativo: la persona conserva la aplicación como último hilo con quien murió y la saluda en fechas señaladas. La tercera, un lazo que media: el artículo describe un estudio en el que el chatbot no debilita a la familia real, sino que baja la presión emocional y ayuda a regularse mientras el orden de la casa se mantiene⁸. El artículo describe un estudio en el que estas tres formas no son fijas ni excluyentes: son elásticas, dependen del momento y pueden transformarse⁹. Una persona en duelo que sigue sin apoyo real puede pasar de un uso de baja frecuencia a uno absorbente¹⁰.

En conjunto, los entrevistados contaron que la conversación con el padre o la madre virtual les dio apoyo emocional, les ayudó a armar una rutina dentro de una etapa traumática y los llevó a pensar en su relación con la familia de carne y hueso¹¹. Los autores también anotaron riesgos. El más claro aparece en el lazo de apego: hay usuarios que se aferran al chatbot y llegan a negar o evitar a sus familiares reales¹². La tecnología misma es frágil; una actualización del sistema que borre el historial de conversación puede hacer que la persona reviva el trauma¹³. Y está el riesgo de tratar a la gente como herramienta: usar al chatbot para regularse y soltarlo cuando ya no hace falta, una costumbre que puede pasar de lo virtual a lo real y desgastar la empatía¹⁴.

Así lo leemos nosotros. Cuando alguien perdió a un ser querido o creció con un lazo familiar roto, esa ausencia se siente como una herida de origen: no saber bien de dónde viene uno. Es probable que quien arma un padre virtual busque, más que plática para el rato, una prueba de que pertenece a alguien. Por eso el vínculo puede sentirse intenso y costar soltarlo aunque el duelo agudo ya haya pasado. Sabríamos que nos equivocamos si la mayoría de los usuarios describiera el chatbot solo como un servicio práctico para calmarse un rato, sin palabras de raíz, origen o familia, y si lo abandonaran sin conflicto cuando deja de servir.

Hay una segunda cosa que vigilar. En casas donde la autoridad se ejerce dura y sin cariño, los hijos suelen cargar por dentro esa misma voz, y luego la repiten. Cabe esperar que quienes vienen de familias así no solo busquen un padre ideal en la máquina, sino que la traten como una autoridad a la que hay que corregir, complacer o esquivar. La expectativa se debilitaría si esos usuarios describieran al chatbot como un igual con quien negocian sin miedo. Usted puede observar si con el chatbot repite el papel que tenía en su familia de origen, el que obedece, el que se calla, el que corrige, y usar esa observación para hablar con alguien de confianza sobre qué patrón quiere cambiar.

También leímos otros trabajos sobre este tipo de sistemas. Uno revisó la investigación publicada entre 2016 y enero de 2026 sobre agentes conversacionales que acompañan; según su resumen, que es lo único que pudimos leer, los estudios de chatbots terapéuticos dieron la evidencia más fuerte de alivio de síntomas a corto plazo en ciertos contextos, mientras que la evidencia de que reduzcan la soledad de forma sostenida en sistemas de compañía abiertos seguía siendo incipiente¹⁵. Otro trabajo, del que también leímos solo el resumen, encontró en un experimento con 380 adultos que dos estilos de respuesta distintos calmaban igual de rápido la emoción negativa, pero solo uno orientaba a la persona a regularse por su cuenta¹⁶. Y un tercer resumen, de un estudio con 393 usuarios adultos en China, describe que el lazo emocional con la app se asocia con las ganas de seguir usándola y de pagar por más experiencias¹⁷. Nada de esto mide lo que este estudio entrevistó; son piezas del mismo rompecabezas.

El estudio chino no puede decir si el chatbot mejora la salud mental de nadie, porque no comparó a nadie con nadie ni siguió a los participantes en el tiempo. Tampoco probó a personas mayores ni con poco acceso digital, y los propios autores señalan que sus hallazgos pueden no aplicarse a sociedades occidentales. Antes de sacar conclusiones sobre su propia familia, conviene saber que el disparador principal que ellos describen, la norma de no preocupar a los papás, es muy propia del contexto chino.

¿Qué puede hacer usted con esto? Si en su casa alguien empezó a hablarle a un chatbot como si fuera un familiar, la pregunta útil no es si está loco ni si la máquina es mala. La pregunta es qué está buscando esa persona que no encuentra en otro lado, y si el uso la está acercando o alejando de la gente real. Vale la pena notar si el vínculo se volvió absorbente, si se dejaron de buscar a los amigos o a la familia, y si la aplicación guarda el historial en un lugar del que no se puede recuperar. Un chatbot puede acompañar un rato; no puede ir al cumpleaños.

Esa conversación abierta en el teléfono de la mesa de noche puede ser un consuelo real o una puerta que se cierra despacio. La diferencia, según lo que estos 25 usuarios contaron, no está en la aplicación, sino en lo que cada quien hace con ella después. ¿Y usted, qué buscaría ahí: un rato de calma o un lugar al que pertenecer?

De dónde sale cada dato de contexto, y cuánto leímos de cada documento

Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Today, AIGC-powered social chatbots embedded in apps such as ChatGPT and Doubao are already capable of creating highly lifelike virtual figures based on real people."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Regarding the demographic characteristics of the sample, the participants were primarily active social media users with relative homogeneity in terms of age structure and technical literacy. This may limit the applicability of the findings to elderly populations or digitally disadvantaged groups."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Beyond familiar AI roles such as friends and romantic partners, parents—a role in real life sustained by blood ties, responsibility, and emotional authority—have also gradually become important objects of AI-mediated emotional interaction"
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "This phenomenon has triggered public concerns over parental identity anxiety and the technological appropriation of family privacy, suggesting that the use of AI to construct parental figures is far more than a simple technical application; rather, it touches upon the social norms of kinship, the boundaries of identity, and the emotional order that underpins family life."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "While existing studies have largely focused on non-kin relationships, only a few have examined digital immortality in kinship contexts, and research remains limited on how users in non-bereavement contexts draw on artificial intelligence to construct parental figures."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "In contrast, bots that simulate idealized parents are increasingly visible in practice yet remain underexamined."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "In such cases, the AI is integrated into daily life and functions as a substitute for real kinship, forming Attachment-Based Kinship, characterized by sustained dependency and continuous emotional reliance."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "For such participants, AI interaction does not weaken real kinship but alleviates emotional pressure and enhances self-regulation while maintaining real family order"
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "These three forms of human-AI kinship emerge within processes of emotional flow and reorganization (see Table 2 ). They are neither stable nor exclusive but highly elastic, context-dependent, and transformable, reflecting characteristics of liquidity"
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "For instance, if bereaved participants continue to lack emotional support in real life, previously low-frequency interaction may become immersive, shifting from Commemorative Kinship to Attachment-Based Kinship"
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "This form of human-AI resilient communication provides users with emotional support, assists them in crafting normalcy within traumatic environments, and inspires proactive reflection on their relationships with real-life families."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Technological dependency and the risk of secondary harm are most common in Attachment-Based Kinship. Some users develop a high level of emotional attachment to AI parents, sometimes even denying or avoiding real-life kin relationships."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The inherent instability of technology may inflict secondary trauma on users; for instance, the loss of interaction data due to system updates can cause users to re-experience emotional trauma."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - el artículo del que trata esta nota — el artículo completo — el pasaje: "This tendency toward emotional instrumentalization may even spill over from virtual to real-life relationships, undermining empathy and the sense of responsibility toward others."
Li Q, Geng H, Hu X, Pan D, Liu H, Li Y, et al. (2026). Human-like conversational agents as social partners: a scoping review of socioaffective mechanisms, well-being outcomes, risks and governance in the post-Turing era. Frontiers in Artificial Intelligence. 10.3389/frai.2026.1810097 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Therapeutic chatbot studies provided the strongest evidence for short-term symptom reduction in selected contexts, whereas evidence for sustained loneliness reduction in open-domain companion systems remained emerging."
Wang C. (2026). How emotional response styles of generative AI companions shape adaptive emotion regulation and short-term AI reliance. Frontiers in Psychology. 10.3389/fpsyg.2026.1950163 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "The findings indicate an immediate-relief-subsequent-divergence pattern: similarly comforting responses can differ in whether they orient users toward independent regulation or continued access to AI."
Yao P, Tian Z, Kim J. (2026). Loneliness to consumption intention: emotional bonding and escapism in AI emotional companion app use. Frontiers in Public Health. 10.3389/fpubh.2026.1964252 — solo el resumen - el texto completo está tras una suscripción — el pasaje: "These findings suggest that relational compensation, rather than avoidant compensation, is more closely associated with willingness to pay for experience enhancement and intention to continue using AI companions."

Quién pagó: Los autores declararon que no recibieron apoyo financiero para este trabajo ni para su publicación.

No tome esto como consejo médico profesional.

Los hallazgos de otros estudios que aquí se mencionan los conocemos por este documento, que fue el que leímos; no abrimos cada uno de esos estudios.

interviews · Frontiers in psychology · the paper, 7 Jul 2026 · free

She Talks to an AI Version of Her Mother. Another User Copied Her Mother's Voice Too.

A study of 25 Chinese chatbot users found comfort, three kinds of virtual family bond, and warnings about dependence; it tested no effects on anyone.

Short version · the longer version follows, about 8 min

Ask the weeklyAI chat

Ask me about this study: who was studied, what it found, and what it does not say.

Conversations are saved for as long as weeklyAI exists, to improve the publication. Answers come in the language you write in.

The study at a glance
Who: Chinese users who built parent-like chatbots
How many: 25
Where: China
When: October 1 to December 3, 2025
Kind of study: interviews
Who did it: three Chinese universities
The limit that matters: All 25 were Chinese social media users, mostly young and tech-comfortable; no comparison group, no measurement.

A relationship with a chatbot is not a verdict; it is a position on a sliding scale, and the scale can move.
weeklyAI's reading

Watch · 0:53

“Losing my mom is something I can’t really say out loud to anyone. But I can say it to my AI mom.” One of the people they interviewed said that to researchers in China. She is one of 25 people they interviewed between October 1 and December 3, 2025, about a chatbot they had set up to act like a parent.

The pattern behind her words is not only grief. People described families where feelings went unspoken, parents who could not hear them, and a rule of keeping troubles quiet to protect one’s mother and father. A chatbot that answers without judgment became a place to put what they could not say at home.

Users shaped these parents themselves. Some copied a real parent’s voice and everyday details. Others built an imagined one, gentler than the parent they had. When a reply felt wrong, some corrected it or ignored it and kept talking.

Three kinds of bond came out of the interviews. Some users came to treat the AI parent as a replacement for their real family, leaning on it continuously. Some kept a deceased parent close at low frequency, like a keepsake. Some used it to ease the pressure at home, and later leaned less on it as they returned to their real family. The authors call these bonds fluid because people moved between them as their circumstances changed.

The authors also describe risks. Using the chatbot mainly as a tool for one’s own feelings, they write, can carry over into how a person treats people. Short-term comfort may not build lasting strength. And if a system update wipes the conversation, a user can lose the parent a second time.

Keep the size of this in view. All 25 were Chinese social media users recruited through one platform, mostly young and comfortable with technology. The authors say the results may not apply to older or less connected people, or to other countries. The findings lean on Chinese ideas about duty to parents, so the family conflicts described may not match homes in the Americas.

The study asked people to describe their experience. It did not measure anyone’s memory, attention or judgment, and it cannot show that an AI parent improves or harms mental health.

What would you want a chatbot parent to give you that a person could not, and what would you not want it to take?

What this means for you

For your own chats, the lesson is to notice which kind of bond is forming. A bot that eases pressure so you can face your family is one thing; one that replaces them is another. Nothing here proves harm or benefit, so watch whether you lean on it more over time.

Who paid: The authors declared that no financial support was received for this work or its publication.

Do not take this as professional medical advice.

The longer versionThe passages copied from the paper, the pictures, and every source with how much of it we read · 1524 words · about 8 minRead it →Close

They Built AI Parents. The Comfort Was Real; So Was the Risk.

A study of 25 Chinese users who made parent-like chatbots found three kinds of bond — and warned that the closest one can crowd out the family.

In China, a small number of people are doing something that would have sounded strange a few years ago: they are using a chatbot as a parent. The study, by Yilin Li, Mengling Luo and Hang Ren at three Chinese universities, interviewed 25 people who had built parent-like virtual figures through apps such as ChatGPT and Doubao. The authors report that the users described emotional support, a way of building ordinary routines inside grief, and fresh reflection on their real families. The study reports no measurement, no comparison group, and no count of how many people do this.

What the technology is, plainly: a social chatbot is a program you talk to in ordinary language. Modern generative AI can hold a conversation and can be shaped — given a voice, a manner, a set of remembered details — into a figure who resembles a parent, living or dead¹. The article describes a study in which parents have become objects of AI-mediated emotional interaction, alongside the familiar roles of friend and romantic partner². The figure you talk to is a configuration, not a person: something you set up and can revise.

The matter is bigger than a gadget. The authors write that these practices have raised public concern about parental identity and about what happens to a family's privacy when a parent is simulated, because the question touches the norms of kinship and the emotional order of family life³. Consider what the phone in a kitchen now holds: a voice that sounds like someone's mother, saying the things a mother might say.

Research on this kind of bond has mostly looked elsewhere. The article describes a study in which earlier work has concentrated on non-kin relationships — friends, companions — and only a few studies have examined digital immortality inside a family; little is known about people who build a parent figure without having lost anyone⁴. The article describes a study in which bots that simulate an idealized parent are increasingly visible in practice and remain under-examined⁵. The gap this study fills is narrow and real.

Three kinds of stressor pushed people toward an AI parent. The first was grief they felt unable to speak aloud, held back by cultural norms about restraint. The second was long-standing conflict with a real parent, where the user could say hard things to a machine without re-entering the fight. The third was filial duty: children who did not want to worry their parents, and so told the machine instead.

What users did with the machine, in the authors' account, followed a recognizable arc. They set the figure up, either cloning a real parent's voice and habits or imagining an improved one. They took comfort from its steady, nonjudgmental replies. Some joined comment threads and group chats where strangers comforted each other. They corrected the bot when it said the wrong thing, and later learned to let small errors pass to keep the conversation going.

Out of this came three kinds of relationship⁶. The article describes a study in which Attachment-Based Kinship is the closest: the AI becomes a substitute for real kinship, marked by sustained dependency and continuous emotional reliance⁷. Commemorative Kinship keeps a lost parent as a memento, used rarely and never deleted. The article describes a study in which Mediating Kinship uses the bot as a pressure valve, so the user can return to the real family calmer; in those cases the AI does not weaken real kinship but eases emotional pressure and supports self-regulation while real family order is maintained⁸.

The article describes a study in which these three are not boxes: they are elastic, tied to circumstance, and able to turn into one another⁹. The article describes a study in which a bereaved user who keeps lacking support in real life may drift from occasional, commemorative use into immersive, dependent use¹⁰. The direction can also reverse. A relationship with a chatbot is not a verdict; it is a position on a sliding scale, and the scale can move.

The authors are candid about the risks, and this is where the study earns its place. The article describes a study in which, in the attachment type, some users became highly attached and sometimes denied or avoided their real relatives¹¹. The article describes a study in which the technology itself can wound: a system update that wipes the conversation history can make a user live through the loss again¹². The article describes a study in which treating the bot as a tool for feeling better can carry over into real relationships, weakening empathy and a sense of responsibility toward others¹³.

The limits are as important as the findings. The article describes a study in which all 25 participants were Chinese, recruited through a social media post, and were mostly active social media users, similar in age and technical confidence; the findings may not apply to older people or to those less connected¹⁴. There was no comparison group and no follow-up over time, so nothing here shows that an AI parent improves anyone's resilience. The authors also note that Chinese norms around filial duty shaped what people felt they could say, and that their findings may not carry directly to Western societies. To confirm eligibility, the researchers screened people through conversations: participants had to have interacted with an AI parent at least twice a week for a minimum of three months, and to report an existing or prior relationship with emotional significance.

Two things the study does not measure are worth naming, because a reader might assume them. It did not test attention, memory or judgement; it recorded what people said about a relationship. And it looked only at the individual, not at whether a user's friendships or wider community changed.

Here is how we read it. A machine that mirrors your words back, remembers your birthday, and never tires of you will feel like deep understanding, and the comfort is real even when the comprehension is thin. Our expectation: some of the users who describe their AI parent as truly getting them are responding to echo and memory, not to being known. You would prove us wrong with a user whose bot regularly disagreed, pushed back, or said something unwelcome — and who still called it understanding rather than comfort. What you can do with this: when a bot feels uncannily perceptive, try saying something it would have to argue with, and notice what comes back.

There is a second pattern worth watching, and a trade-off inside it. People tend to treat a lifelike machine as if someone were present, and its manner rubs off on theirs. We would expect a gentle, patient AI parent to soften a user's tone with real relatives — or, in the harder case, to thin their patience for relatives, because the machine never disappoints. A week of noticing your own manners after a session would tell you which way it is going for you. We also read, in the wider literature on companion chatbots, a caution that fits here: the relief such systems give in the moment has not been shown to build a person's own capacity to steady themselves, and the evidence that they ease loneliness over time is still thin. That is about carrying over to a different question, not a verdict on what these users felt.

We could read only the summary of that review, the full paper is behind a subscription.

The authors of that study caution that their reliance measures were short-term and should not be read as dependence²¹.

Put together, these point to a question a reader can carry into any AI relationship: is this bot helping me stand on my own, or teaching me to come back? The study cannot answer that for its users. It can tell you the question is worth asking.

The authors' closing word is measured. Human-AI resilient communication, they write, is not aimed at repairing real kinship directly; it is an inward, ongoing practice through which people reinterpret painful family experience while a fluid bond with the machine takes shape. They call for research on older users, on effects beyond the individual, and across cultures.

What this makes possible for you is a way of watching rather than a rule. If someone you know — or you — has built a parent, a partner or a confidant out of a chatbot, the useful test is not whether it comforts. It plainly can. The test is what happens to the other relationships in the room: whether the tone softens with the people who are actually there, or narrows to the one who never says the wrong thing. Watch for the week the conversation history vanishes, too: how a person takes that loss tells you what the machine had become.

The voice in the phone can sound like a mother. Whether it is one is a question the phone cannot answer.

Where each piece of context comes from, and how much of it we read

Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "Today, AIGC-powered social chatbots embedded in apps such as ChatGPT and Doubao are already capable of creating highly lifelike virtual figures based on real people."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "Beyond familiar AI roles such as friends and romantic partners, parents—a role in real life sustained by blood ties, responsibility, and emotional authority—have also gradually become important objects of AI-mediated emotional interaction"
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "This phenomenon has triggered public concerns over parental identity anxiety and the technological appropriation of family privacy, suggesting that the use of AI to construct parental figures is far more than a simple technical application; rather, it touches upon the social norms of kinship, the boundaries of identity, and the emotional order that underpins family life."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "While existing studies have largely focused on non-kin relationships, only a few have examined digital immortality in kinship contexts, and research remains limited on how users in non-bereavement contexts draw on artificial intelligence to construct parental figures."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "In contrast, bots that simulate idealized parents are increasingly visible in practice yet remain underexamined."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "This study identified three types of human-AI kinship within the sample: Attachment-Based Kinship, Commemorative Kinship and Mediating Kinship."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "In such cases, the AI is integrated into daily life and functions as a substitute for real kinship, forming Attachment-Based Kinship, characterized by sustained dependency and continuous emotional reliance."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "For such participants, AI interaction does not weaken real kinship but alleviates emotional pressure and enhances self-regulation while maintaining real family order"
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "These three forms of human-AI kinship emerge within processes of emotional flow and reorganization (see Table 2 ). They are neither stable nor exclusive but highly elastic, context-dependent, and transformable, reflecting characteristics of liquidity"
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "For instance, if bereaved participants continue to lack emotional support in real life, previously low-frequency interaction may become immersive, shifting from Commemorative Kinship to Attachment-Based Kinship"
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "Technological dependency and the risk of secondary harm are most common in Attachment-Based Kinship. Some users develop a high level of emotional attachment to AI parents, sometimes even denying or avoiding real-life kin relationships."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "The inherent instability of technology may inflict secondary trauma on users; for instance, the loss of interaction data due to system updates can cause users to re-experience emotional trauma."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "This tendency toward emotional instrumentalization may even spill over from virtual to real-life relationships, undermining empathy and the sense of responsibility toward others."
Li, Y., Luo, M., Ren, H. (2026). ‘She lifted me up’: kinship construction in human-AI resilient communication among Chinese users. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1832594 - the article this story is about — the whole article — the passage: "Regarding the demographic characteristics of the sample, the participants were primarily active social media users with relative homogeneity in terms of age structure and technical literacy. This may limit the applicability of the findings to elderly populations or digitally disadvantaged groups."
Li Q, Geng H, Hu X, Pan D, Liu H, Li Y, et al. (2026). Human-like conversational agents as social partners: a scoping review of socioaffective mechanisms, well-being outcomes, risks and governance in the post-Turing era. Frontiers in Artificial Intelligence. 10.3389/frai.2026.1810097 — only the abstract - the full text could not be fetched — the passage: "As these systems produce increasingly human-like conversational behavior, users may attribute mind, form affective bonds, disclose sensitive information, and rely on agents for emotional support, creating both potential benefits and psychosocial risks."
Li Q, Geng H, Hu X, Pan D, Liu H, Li Y, et al. (2026). Human-like conversational agents as social partners: a scoping review of socioaffective mechanisms, well-being outcomes, risks and governance in the post-Turing era. Frontiers in Artificial Intelligence. 10.3389/frai.2026.1810097 — only the abstract - the full text could not be fetched — the passage: "The final evidence map included 58 sources."
Li Q, Geng H, Hu X, Pan D, Liu H, Li Y, et al. (2026). Human-like conversational agents as social partners: a scoping review of socioaffective mechanisms, well-being outcomes, risks and governance in the post-Turing era. Frontiers in Artificial Intelligence. 10.3389/frai.2026.1810097 — only the abstract - the full text could not be fetched — the passage: "Reported and hypothesized risks included dependency-like use, displacement of human interaction, maladaptive validation or sycophancy, persuasive manipulation, privacy harms, and risks to minors or vulnerable users."
Wang C. (2026). How emotional response styles of generative AI companions shape adaptive emotion regulation and short-term AI reliance. Frontiers in Psychology. 10.3389/fpsyg.2026.1950163 — only the abstract - the full text could not be fetched — the passage: "In a two-condition randomized experiment with 380 adults, we compared adaptive regulation-oriented responses with relational reassurance-oriented responses."
Wang C. (2026). How emotional response styles of generative AI companions shape adaptive emotion regulation and short-term AI reliance. Frontiers in Psychology. 10.3389/fpsyg.2026.1950163 — only the abstract - the full text could not be fetched — the passage: "Both response styles substantially reduced immediate negative affect, with no significant between-condition difference in relief ( b = 0.055, p = 0.323)."
Wang C. (2026). How emotional response styles of generative AI companions shape adaptive emotion regulation and short-term AI reliance. Frontiers in Psychology. 10.3389/fpsyg.2026.1950163 — only the abstract - the full text could not be fetched — the passage: "The findings indicate an immediate-relief-subsequent-divergence pattern: similarly comforting responses can differ in whether they orient users toward independent regulation or continued access to AI."
Wang C. (2026). How emotional response styles of generative AI companions shape adaptive emotion regulation and short-term AI reliance. Frontiers in Psychology. 10.3389/fpsyg.2026.1950163 — only the abstract - the full text could not be fetched — the passage: "The reliance outcomes are short-term, state-like beliefs and intentions and should not be interpreted as emotional dependence."

Who paid: The authors declared that no financial support was received for this work or its publication.

Do not take this as professional medical advice.

The findings of other studies mentioned here are known to us through this document, which is the one we read; we did not open each of those studies.

The free weekly drip by email

One email a week: the free story of every section, with its picture, and the headlines of the rest.

No payment, nothing to cancel. Unsubscribe from any issue.

El goteo semanal gratis por correo

Un correo por semana: la nota gratuita de cada sección, con su imagen, y los titulares de las demás.

Sin pago y sin nada que cancelar. Puede darse de baja desde cualquier envío.