"It Feels Like My Little Online Therapist": What 21 Teenagers Said About AI and Self-Injury

US · CanadaAmérica Latina

Weekly AI Watch

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

Week of 30 October 2026EN | ESPast editions15,230 papers found · 14,455 read · 4,301 belonged · 46 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 30 de octubre de 2026EN | ESEdiciones anteriores15,230 estudios encontrados · 14,455 leídos · 4,301 correspondían · 46 notas pasaron todas las comprobaciones

weeklyAI · Week of 30 October 2026weeklyAI · Semana del 30 de octubre de 2026

← Your child← Sus hijos

interviews · Frontiers in digital health · la publicación, 26 ago 2026 · gratis

"Le escribo a ChatGPT… siento que es mi terapeuta en línea"

En entrevistas con 21 adolescentes de Estados Unidos que se lastiman a sí mismos, varios contaron que usan chatbots de IA para hablar de lo que no le dirían a una persona. También chocaron con filtros de seguridad que cortaban la conversación.

Versión breve · la versión detallada sigue, unos 6 min

Pregunte al chat de weeklyAI

Pregúnteme sobre la IA en su vida: le señalo las notas que weeklyAI ha publicado sobre el tema, esta semana y antes.

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: Adolescentes que se lastiman a sí mismos
Cuántos: 21 adolescentes
Dónde: Estados Unidos
Cuándo: Entre mayo de 2024 y abril de 2025
Tipo de estudio: interviews
Quién lo hizo: Universidad Northwestern, con una investigadora principal que además es trabajadora social clínica
El límite que importa: Quienes contestaron un cuestionario en línea quizá estén más atentos a su salud mental y más abiertos a lo digital que la mayoría.

Lo que ahora sabemos es que ese programa no reemplaza a nadie, que a veces contesta y a veces se cierra, y que el joven al otro lado está pidiendo, casi siempre, un lugar donde no lo juzguen.
Lectura de weeklyAI

Ver · 0:28

"I message ChatGPT… it feels like it's my little online therapist, and that helps", dijo uno de los participantes. Le resultaba más cómodo que contárselo a alguien conocido.

Los investigadores entrevistaron a 21 adolescentes de entre 14 y 18 años que se lastimaban a sí mismos. En el mes anterior lo habían hecho en promedio unos cinco días, y algunos hasta veinte. Los reclutaron entre mayo de 2024 y abril de 2025 a través de la página de una organización nacional de salud mental. Las entrevistas fueron a distancia.

Detrás de esa frase hay un patrón. Los adolescentes acudían a la IA para ordenar lo que sentían, para preguntar cosas que les daban vergüenza preguntarle a una persona y para hablar cuando no había nadie más. Varios lo contaron con cierta pena, como si confesar algo raro. "Es casi más fácil abrirse porque sabes que no es humano", explicó otra participante.

El estudio no midió si eso les ayudó o les hizo daño. No comparó a nadie con nadie ni siguió a los adolescentes en el tiempo. Lo que recoge son sus palabras.

Las mismas herramientas les fallaron. Cuando querían hablar de lastimarse, los filtros de seguridad respondían que no hablarían del tema. "A veces necesito hablar de eso", dijo uno. Cuando sí contestaban, el consejo era genérico: "confía en un amigo o escribe un diario". Otros recibieron respuestas equivocadas.

Puede que estén más atentos a su salud mental y más abiertos a lo digital que la mayoría. Y lo que describieron son las herramientas de 2024 y principios de 2025, que no son las de hoy.

Lo que sí es concreto: los adolescentes ya están teniendo estas conversaciones, muchas veces sin que nadie en casa lo sepa. Preguntarles qué le piden a la IA y qué les responde es un lugar razonable para empezar.

¿Qué querría su hijo que una IA le dijera cuando menciona que se lastima, y qué no debería decirle nunca?

Qué significa para usted

Lo que esos adolescentes describieron fueron herramientas de 2024 y principios de 2025, y el estudio no midió si los chatbots ayudaron o dañaron a nadie. En casa puede preguntar qué le cuenta su hijo a la inteligencia artificial cuando está mal y qué le contesta ella. Si la respuesta lo deja sin palabras, ese es el tema que conviene llevar a un profesional.

Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943

Quién pagó: El trabajo recibió apoyo de una subvención del Instituto Nacional de Salud Mental (K01MH131898); el artículo no indica si el financiador tuvo algún papel, y no se mencionan donaciones de equipos ni programas.

No tome esto como consejo médico profesional.

Versión detalladaLos pasajes copiados del artículo, las ilustraciones y cada fuente con cuánto leímos de ella · 1206 palabras · unos 6 minLeerla →Cerrar

Lo que 21 adolescentes contaron sobre la autolesión y sus teléfonos

Un estudio escuchó a jóvenes que se lastiman. No midió si la tecnología ayuda o daña; preguntó cómo la usan y qué les falta.

Un teléfono en la mano a las once de la noche, y al otro lado un programa que contesta. El estudio que leemos esta semana se hizo justamente para entender esa escena: cómo adolescentes que se lastiman a propósito usan la tecnología para sobrellevarlo, y qué esperarían de ella. Los autores entrevistaron a 21 jóvenes de entre 14 y 18 años, con una edad promedio de casi 17, que habían hecho esto en dos o más días del mes anterior. Las entrevistas fueron a distancia, entre mayo de 2024 y abril de 2025, y los reclutaron por el sitio de una organización nacional de salud mental. El trabajo lo firma un equipo de la Universidad Northwestern, con una investigadora principal que además es trabajadora social clínica, y lo pagó una beca del Instituto Nacional de Salud Mental de Estados Unidos.

Lo primero que hay que decir es lo que el estudio no hizo. No comparó nada, no midió si el uso de estas herramientas mejora o empeora la salud de nadie, y no evaluó ninguna app ni ningún sistema en particular. Reporta lo que estos jóvenes dijeron en una entrevista. Con ese tamaño y esa puerta de entrada, la muestra no representa a todos los adolescentes que se lastiman: quienes contestaron un cuestionario en línea quizá estén más atentos a su salud mental y más abiertos a lo digital que la mayoría¹.

Conviene poner el asunto en su tamaño. Alrededor del 17 por ciento de los adolescentes se lastima a propósito, según las cifras que el propio artículo cita, y la mayoría no recibe tratamiento formal: el estigma, el costo y la falta de terapeutas dejan ese hueco abierto²³. Por eso los autores buscaban alternativas accesibles, no un juguete nuevo.

En las entrevistas, los jóvenes describieron la autolesión como una forma de aguantar, no como un capricho: algo que aparece cuando la escuela, la familia o la cultura aprietan demasiado. Muchos dijeron estar conscientes de que les hace daño y, aun así, sentirse divididos sobre dejarlo. Y ahí empieza el problema práctico: hablar de esto con un adulto les cuesta. Un participante lo explicó con claridad: pedir ayuda a un terapeuta siendo menor de edad choca con la obligación de avisar a los padres, y eso lo frenó⁴. Otro dijo que casi nunca lo ha hablado en persona, aunque la gente cercana lo sepa⁵.

Con ese panorama, la tecnología aparece como refugio. Los participantes contaron que usan plataformas de música, apps para escribir un diario o registrar el ánimo, comunidades en línea y sistemas de inteligencia artificial como ChatGPT para regular lo que sienten, pensar en lo que les pasa y sacar temas delicados⁶. Una joven lo dijo sin rodeos: le escribe a ChatGPT y siente que es su terapeuta de bolsillo⁷. Otra explicó por qué le resulta más fácil: sabe que no es humano⁸. Y una tercera dio la razón de fondo: puede desahogarse sin culpa, porque no siente que esté fastidiando a nadie⁹.

Los jóvenes describieron esa conversación como una charla que devuelve algo, a diferencia de un diario que se queda callado: dijeron que la inteligencia artificial les hace preguntas y recuerda detalles. Eso es lo que los jóvenes describieron como una charla que devuelve algo, a diferencia de un diario que se queda callado. Los propios autores señalan que las entrevistas se hicieron en meses en que estas herramientas cambiaban rápido, de modo que lo descrito puede no coincidir con lo que un joven usa hoy.

La misma herramienta que consuela también falla, y los jóvenes lo dijeron. Cuando quisieron hablar de autolesión, los filtros de seguridad cortaron la conversación: el sistema contestó que de eso no se habla, justo cuando necesitaban hablarlo¹³. Cuando sí respondió, el consejo salió genérico: confía en un amigo, escribe en tu diario, lo mismo que ya habían oído mil veces¹⁴. Y a veces se equivocó de plano¹⁵. Los participantes notaron además una trampa en las comunidades en línea: pueden acompañar un rato, pero con el tiempo normalizan y hasta vuelven competitiva la conducta, con la sensación de que uno "no se lastima lo suficiente"¹⁶¹⁷.

Aquí es donde nosotros leemos un patrón que conviene tener a la vista. Cuando alguien revela algo íntimo, queda expuesto: teme que esa confianza se use en su contra, y prefiere callar. Un programa que no juzga y no puede avisar a nadie se vuelve entonces la puerta más barata. Los jóvenes del estudio contaron que les resulta más fácil abrirse con una máquina porque saben que no es humana y no sienten que estén fastidiando a nadie. Nos equivocaríamos si los jóvenes de este estudio hubieran descrito conversaciones abiertas con adultos cercanos sin miedo al juicio, o si contarle algo a un adulto los hubiera acercado a pedir ayuda en vez de alejarlos. No fue lo que contaron.

De ahí sale algo concreto para cuando llegue ese momento: cuide la reacción antes que el contenido. Escuche sin dramatizar, pregunte qué necesita antes de opinar, y no tome como traición que prefiera hablar con una máquina; tómelo como pista de que hace falta un espacio humano seguro. Los propios autores lo apuntan: responder con empatía, interés y curiosidad, sin confundir la autolesión con ideas suicidas ni enfocarse solo en la conducta, es lo que evita que el joven se cierre¹⁸. La otra mitad es la privacidad: varios dijeron que usarían una app si fuera un espacio confidencial que no etiquete el contenido como autolesión, y propusieron entradas con Face ID, doble verificación, pantallas de salida rápida y hasta nombres de disfraz¹⁹.

Quedan límites que un lector podría pasar por alto. La muestra fue en su mayoría blanca y femenina, y se excluyó a jóvenes con riesgo suicida alto o con ciertos diagnósticos. Tampoco están los que no han contado nada en casa: para proteger la confidencialidad no se pidió consentimiento de los padres, y a cada joven se le pidió el nombre y el teléfono de un adulto de confianza al que se pudiera llamar si aparecía riesgo de suicidio. Y las entrevistas se hicieron en meses en que las herramientas de inteligencia artificial cambiaban rápido, de modo que lo descrito puede no coincidir con lo que su hijo usa hoy.

Así lo leemos: lo que estos jóvenes buscan no es un sustituto de las personas sino un primer paso, y un chatbot que contesta y recuerda detalles puede sentirse como compañía sin serlo del todo. Usted puede comprobarlo en su casa: pregunte con quién habla de lo que le preocupa, en persona y en línea, y ofrezca compañía concreta sin interrogatorios. Si la respuesta es "con nadie" o "con una app", ya tiene por dónde empezar.

Volvamos al teléfono de las once de la noche. Lo que ahora sabemos es que ese programa no reemplaza a nadie, que a veces contesta y a veces se cierra, y que el joven al otro lado está pidiendo, casi siempre, un lugar donde no lo juzguen. ¿Qué necesita su hijo o hija para que ese lugar pueda ser usted?

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

Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "First, the sample consisted of adolescents who voluntarily completed an online mental health screener. Compared with adolescents who do not engage with such resources, participants may have had greater awareness of their mental health concerns, been more motivated to seek support, and more receptive to digital mental health tools."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Nonsuicidal self-injury (NSSI)—hereafter referred to as “self-injury”—is reported by approximately 17% of adolescents and is associated with a range of co-occurring mental health challenges and elevated risk of suicide ( 1 , 2 )."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Yet, most adolescents that engage in self-injury do not receive formal treatment. Stigma surrounding mental health concerns, and self-injury in particular, as well as structural barriers such as treatment costs and limited access to providers, contribute to this gap and highlight a need for alternative forms of intervention."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "PID14 said “It's especially hard as a younger person because you have this barrier with your parents, I guess, because there are certain things that a therapist would be obligated to tell your parents, and I think self-harm is one of those, which makes it hard for under-age people to get help.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "PID16 said “[Stigma] is so incredibly difficult - even though everyone around me in person knows that I self-injure, I've barely ever actually talked about it. It's just a very difficult thing to talk about in person.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Participants reported using a range of tools, including music streaming platforms, journaling and mood-tracking apps, online communities, and AI systems such as ChatGPT to regulate emotions, reflect on their experiences, seek advice, and discuss sensitive topics."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "PID11 noted, “I message ChatGPT… it feels like it's my little online therapist, and that helps,” emphasizing that talking to an AI could feel more comfortable than confiding in someone they knew."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Emphasizing the absence of interpersonal pressure when talking with AI agents, PID16 explained, “It's almost easier to open up to, because you know it's not human.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "PID12 shared this sentiment, “You can rant guilt free to a bot because you're not being annoying. But it's like, if it's another person, you can't talk to a person to this extent or expect them to be there to this extent.”"
Maheux AJ, Boeldt D, Akre-Bhide S, Flannery J, Paige A, Villella G, et al. (2026). How US Youth Use AI Chatbots: Conversation Patterns From Naturalistic Keystroke Observations. Journal of Medical Internet Research. 10.2196/95819 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "This study used naturalistic behavioral observations of keystrokes typed into GenAI mobile apps to characterize the topics youth discuss with chatbots, describing themes by age and app, word count by theme (a proxy for quantity of engagement), and co-occurring themes."
Maheux AJ, Boeldt D, Akre-Bhide S, Flannery J, Paige A, Villella G, et al. (2026). How US Youth Use AI Chatbots: Conversation Patterns From Naturalistic Keystroke Observations. Journal of Medical Internet Research. 10.2196/95819 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Participants were 3363 US youth aged 1-17 years using the Aura parental monitoring app, a commercially available service that guardians install on children's mobile devices for digital safety."
Maheux AJ, Boeldt D, Akre-Bhide S, Flannery J, Paige A, Villella G, et al. (2026). How US Youth Use AI Chatbots: Conversation Patterns From Naturalistic Keystroke Observations. Journal of Medical Internet Research. 10.2196/95819 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Theme prevalence varied by age: tool use was more common among older adolescents, whereas violence and sexual or romantic role-play were relatively more common among younger users."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "PID11 described how safety restrictions often blocked important conversations they hoped to have: “I’ve used a few different AI therapy-specific models. But they’re like, a little bit more limited, or if I want to talk about maybe self-injury – it’ll just be like, you can't, we’re not talking about this – like, it doesn't want to continue that conversation, and sometimes I need to talk about it.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "PID18 explained, “If I was having a particular struggle moment, a very intense moment, I wanted specific answers to my specific situation — [But] it was also like, ‘Confide in a friend or journal,’ and a lot of the other very general mental health reminders that I’ve been surrounded by in communities.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "As PID19 put it, “I’ve just seen the Google AI be blatantly wrong sometimes.” While most participants recognized the fallibility of these systems when it came to factual information, they were somewhat less critical when receiving emotional validation or situational advice."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "PID21 reflected that visiting communities helped in the short term but had a negative effect in the long term: “Talking about it is keeping it from getting better, and encouraging other people to do it too. So I think things are just band-aids over a bullet wound.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "PID16 said “A lot of issues like self-harm, especially online, can even be competitive in online spaces. If you see someone who does it more than you, or has more severe injuries than you, you think, ‘wow, I’m not doing it right,’ or ‘I can't even hurt myself good enough’”."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Our participants emphasized the importance of responding to disclosure with empathy, concern, and curiosity about the unique ways the behavior may present for them. Misunderstandings or assumptions about self-injury - particularly those that associate self-injury with suicide or that focus on the behavior alone - can dissuade adolescents from seeking future help and increase shame."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Several noted that they would feel more inclined and willing to log behaviors if a future app provided a “confidential space” and avoided explicitly labeling content as “self-injury,” making it feel less stigmatizing and risky."

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 digital health · the paper, 26 Aug 2026 · free

"It Feels Like My Little Online Therapist": What 21 Teenagers Said About AI and Self-Injury

In interviews, adolescents who self-injure described using ChatGPT for support. They also hit filters that refused to discuss self-injury, and the study cannot show whether any of it helps.

Short version · the longer version follows, about 5 min

Ask the weeklyAI chat

Ask me about AI in your life: I point you to the stories weeklyAI has published on it, this week and before.

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: adolescents who self-injure
How many: 21
Where: United States
When: May 2024 to April 2025
Kind of study: interviews
Who did it: Northwestern University and others
The limit that matters: Only 21 teenagers, recruited through one mental health website; no measure of whether tools helped or harmed.

The attraction is not that the machine is wise. It is that the machine is always there and never flinches.
weeklyAI's reading

Watch · 0:33

"I message ChatGPT," one teenager said. "It feels like it's my little online therapist, and that helps."

That sentence comes from a study of 21 adolescents, ages 14 to 18, who had self-injured at least twice in the past month. Researchers at Northwestern University interviewed them between May 2024 and April 2025. They were recruited through the website of Mental Health America, a national mental health nonprofit.

Behind that sentence was a pattern. Teenagers turned to AI to sort out feelings, to ask questions too embarrassing to ask a person, and to talk when no one else was available. "It's almost easier to open up to, because you know it's not human," another participant said.

The same teenagers ran into walls. Safety filters shut down conversations about self-injury. "It'll just be like, you can't, we're not talking about this," one said, "and sometimes I need to talk about it." Others got advice they had heard before, like "confide in a friend or journal," when they wanted answers to their specific situation. Some caught the AI being plainly wrong.

The study reports what these 21 teenagers said. It did not measure whether chatbots helped or harmed their mental health over time. All 21 volunteered through an online mental health screener, so they may be more aware of digital tools than most teenagers. Your child's experience may differ.

The interviews ran from May 2024 to April 2025, a stretch when AI tools and the public debate around them changed quickly. What these teenagers described may not match the apps your child uses today.

The teenagers also used music apps, journaling and mood-tracking apps, and online communities. They said privacy, discretion and feeling safe mattered most in whether they would use an app for self-injury. They suggested lock screens and apps that could pass as something else. Online communities could connect them with others, but some said those spaces made self-injury worse.

What would your child want an AI to say when they mention self-injury, and what should it never say?

What this means for you

These teenagers described using apps, online communities and chatbots like ChatGPT to cope, and named privacy, canned answers and triggering content as their main worries. The study cannot say whether any of it helped. Watch whether your child's tools answer honestly about self-injury or shut the subject down, and ask a professional about the rest.

Who paid: The work was supported by a grant from the National Institute of Mental Health (K01MH131898); the article does not say whether the funder had any role, and no equipment or software donations are mentioned.

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 · 982 words · about 5 minRead it →Close

# When a Teenager Tells a Chatbot What They Won't Tell You

Twenty-one adolescents who self-injure described using apps, online communities and chatbots like ChatGPT to cope. They also described what those tools get wrong.

The study asked 21 adolescents who self-injure about the technology they use to manage it. They described using apps, online communities and chatbots like ChatGPT to cope, and they described what those tools get wrong.

A study in Frontiers in Digital Health asked 21 adolescents, ages 14 to 18, who were US citizens or residents, about their self-injury and about the technology they use to manage it. All had self-injured on at least two days in the past month. The interviews ran from May 2024 to April 2025, and the researchers analyzed what the teenagers said by grouping their accounts into common themes.

The study reports no measured effect of any tool on anyone's mental health. It is a record of what these 21 young people said about their own lives, and the authors are clear about that.

What they said was that self-injury worked as a coping strategy inside larger distress: school pressure, family and cultural expectations, and social media they described as hard to put down. Many were unsure they wanted to stop. Most had been in therapy at some point, and most of those had told their provider about their self-injury.

The article describes a study in which nonsuicidal self-injury is reported by roughly 17 percent of adolescents, and most of them never receive formal treatment¹². The article describes a study in which the authors point to stigma, cost and scarce providers as the reasons². That gap is the background against which a phone becomes a plausible source of help.

Here is how the tools were used, in the teenagers' own accounts. Music streaming to shift a mood or outlast an urge, journaling apps to see progress, mood trackers to notice patterns, and online communities to find people who understood³. A chatbot was something else again: someone to talk to. One participant called it "my little online therapist"⁴. Another said it was easier to open up "because you know it's not human"⁵. A third described venting "guilt free" to a bot, because a bot cannot be burdened⁶.

Participants said they turned to a chatbot to talk about things they felt they could not bother other people with. Participants valued the chatbot for being available at any hour and for not being a person. One participant said it felt easier to open up to it because it was not human.

It also fails in ways the teenagers could name. Safety filters shut conversations down: one participant was told, in effect, we're not talking about this, at the moment they most needed to⁷. Even when the system answered, the advice was generic: confide in a friend, journal⁸. One said the search assistant was "blatantly wrong sometimes"⁹. And the researchers note that participants were less critical of wrong answers when the answer was emotionally comforting⁹.

Online communities carried their own double edge. They offered solidarity, and they could make things worse. One participant described self-harm content online as competitive, and said those spaces became a loop where no one was focusing on recovery¹⁰. Another called those spaces "band-aids over a bullet wound"¹¹. Music, too, could validate or intensify, depending on the song³.

Age was its own barrier. Teenagers feared that disclosing to a therapist would trigger a report to their parents, and that fear kept them quiet¹². Stigma made even ordinary conversation difficult: one participant said everyone around them knew, and they had barely ever spoken of it aloud¹³.

Privacy was the condition of use. Participants said they would log more honestly in an app that offered a confidential space and avoided the words "self-injury" on the screen¹⁴. They suggested face recognition, two-factor login, a quick-exit screen, even a disguise.

The limits are real. Twenty-one people, recruited through one mental health nonprofit's website, may be more aware of their mental health and more open to digital tools than most teenagers¹⁵. Most were White, and about two-thirds were female. Interviews spanned eleven months during which AI tools and the public argument about them changed quickly. Teenagers at the highest suicide risk were excluded from the study, and parental consent was waived because many adolescents have not told their parents about their self-injury.

Here is how we read it. A teenager who says the chatbot "asks questions" and "remembers details" may be receiving a mirror, not a mentor. The comfort is real; whether it lasts past the novelty is not something this study can tell you. Watch for it: does the conversation ever push back, disagree, tell your child something they did not already feel? If it only reflects and soothes, the calm may be thin.

We also notice what the blocked conversations imply. When a chatbot refuses a topic, that is a design decision made by a company, not a judgment about your child. Those refusals are likely to become more common as the industry faces lawsuits and regulation, and a teenager who hits that wall may look for a place with fewer guardrails. Line up the human fallback before the wall arrives.

Our advice to caregivers, drawn from what the authors themselves conclude, is to respond to a disclosure with empathy and curiosity rather than alarm¹⁶. The researchers warn that conflating self-injury with suicide, or focusing only on stopping the behavior, can drive a teenager further into silence and shame¹⁶. Ask what they type that they cannot say out loud. Ask what the machine says back.

The study cannot say whether the comfort these tools offer lasts. The researchers say caregivers should ask about adolescents' technology use and try to understand how their online worlds relate to their mental health. What would you want your child to hear from you, before they type it to a machine?

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

Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "Nonsuicidal self-injury (NSSI)—hereafter referred to as “self-injury”—is reported by approximately 17% of adolescents and is associated with a range of co-occurring mental health challenges and elevated risk of suicide ( 1 , 2 )."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "Yet, most adolescents that engage in self-injury do not receive formal treatment. Stigma surrounding mental health concerns, and self-injury in particular, as well as structural barriers such as treatment costs and limited access to providers, contribute to this gap and highlight a need for alternative forms of intervention."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "Participants reported using a range of tools, including music streaming platforms, journaling and mood-tracking apps, online communities, and AI systems such as ChatGPT to regulate emotions, reflect on their experiences, seek advice, and discuss sensitive topics."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "PID11 noted, “I message ChatGPT… it feels like it's my little online therapist, and that helps,” emphasizing that talking to an AI could feel more comfortable than confiding in someone they knew."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "Emphasizing the absence of interpersonal pressure when talking with AI agents, PID16 explained, “It's almost easier to open up to, because you know it's not human.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "PID12 shared this sentiment, “You can rant guilt free to a bot because you're not being annoying. But it's like, if it's another person, you can't talk to a person to this extent or expect them to be there to this extent.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "PID11 described how safety restrictions often blocked important conversations they hoped to have: “I’ve used a few different AI therapy-specific models. But they’re like, a little bit more limited, or if I want to talk about maybe self-injury – it’ll just be like, you can't, we’re not talking about this – like, it doesn't want to continue that conversation, and sometimes I need to talk about it.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "PID18 explained, “If I was having a particular struggle moment, a very intense moment, I wanted specific answers to my specific situation — [But] it was also like, ‘Confide in a friend or journal,’ and a lot of the other very general mental health reminders that I’ve been surrounded by in communities.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "As PID19 put it, “I’ve just seen the Google AI be blatantly wrong sometimes.” While most participants recognized the fallibility of these systems when it came to factual information, they were somewhat less critical when receiving emotional validation or situational advice."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "PID16 said “A lot of issues like self-harm, especially online, can even be competitive in online spaces. If you see someone who does it more than you, or has more severe injuries than you, you think, ‘wow, I’m not doing it right,’ or ‘I can't even hurt myself good enough’”."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "PID21 reflected that visiting communities helped in the short term but had a negative effect in the long term: “Talking about it is keeping it from getting better, and encouraging other people to do it too. So I think things are just band-aids over a bullet wound.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "PID14 said “It's especially hard as a younger person because you have this barrier with your parents, I guess, because there are certain things that a therapist would be obligated to tell your parents, and I think self-harm is one of those, which makes it hard for under-age people to get help.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "PID16 said “[Stigma] is so incredibly difficult - even though everyone around me in person knows that I self-injure, I've barely ever actually talked about it. It's just a very difficult thing to talk about in person.”"
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "Several noted that they would feel more inclined and willing to log behaviors if a future app provided a “confidential space” and avoided explicitly labeling content as “self-injury,” making it feel less stigmatizing and risky."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "First, the sample consisted of adolescents who voluntarily completed an online mental health screener. Compared with adolescents who do not engage with such resources, participants may have had greater awareness of their mental health concerns, been more motivated to seek support, and more receptive to digital mental health tools."
Ramirez, I., Barnas, O., Mohr, D. C. et al. (2026). Navigating self-injury in a digital world: adolescents’ perspectives on coping, help-seeking, and technology. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1905943 - the article this story is about — the whole article — the passage: "Our participants emphasized the importance of responding to disclosure with empathy, concern, and curiosity about the unique ways the behavior may present for them. Misunderstandings or assumptions about self-injury - particularly those that associate self-injury with suicide or that focus on the behavior alone - can dissuade adolescents from seeking future help and increase shame."

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.