A New Emotional Anchor for the Youth? Confiding in AI with Their Concerns — 28% of Young Users Seek Advice, Challenges Beyond Convenience

At night, it's a bit late to contact a friend. It's hard to explain to family. I'm not even sure if it's something worth consulting a professional about.

At such times, I open my smartphone and pour my feelings into a chatbot. The AI I used for research and writing has somehow become a "listener for my worries."

A survey indicating this change was reported in Germany. According to an article published by BUSINESS-PANORAMA on November 2, 2026, 28% of young chatbot users use AI for personal conversations seeking emotional support, questions about health and treatment, and worries.

However, this figure does not mean "28% of young people are receiving treatment from AI." It could include anything from voicing minor anxieties to consulting about serious issues.

The survey raises not only the question of whether AI can replace humans but also what kind of support and safety measures are needed in an era where people turn to AI before consulting others.

28% of young users engage in personal conversations with AI

According to reports, the survey was commissioned by the German Consumer Federation "vzbv" and conducted by the public opinion research institute Forsa. It targeted 574 people aged 14 to 25 and is described as a representative online survey.

93% of respondents said they had used an AI chatbot in the past 12 months. Among chatbot users, 28% also use AI for personal consultations.

It's important to note the base number for these figures. 93% represents the proportion of respondents with usage experience, while 28% is the proportion of personal conversations among chatbot users.

Furthermore, the survey includes not only minors but also adults in their twenties. Rephrasing it as "about 30% of children consult AI about their worries" would deviate from the original information.

The survey report does not provide details on the questionnaire, usage frequency, or breakdown of services used. It is possible that both people who consult daily for long hours and those who consulted only once in the past are counted as users.

What can be gleaned from this survey is that AI is entering the realm of personal conversations for the younger generation. It is not a survey that directly indicates the spread of dependency or therapeutic effects.

Health, diet, friendships—consultations are close to everyday life

What are people discussing with chatbots in personal conversations?

According to the report, 55% of the topics were related to health, 48% to nutrition and eating behavior, 47% to worries and problems, and 40% to love and friendships. Additionally, 30% used it when they felt lonely or had no one to talk to.

These percentages are among those who engage in personal consultations, not the overall percentage of survey respondents. Also, the numbers do not add up to 100%.

What is noteworthy is the wide range of consultation topics.

Health issues, eating habits, misunderstandings with friends, uncertainties in love. All are everyday topics, yet sometimes difficult to bring up with others. Consulting AI is not necessarily an action limited to those "not connected to anyone."

Even with people around, there are things that can't be discussed immediately. Sometimes the words to explain one's worries haven't been found yet. It can be thought that chatbots, which are always ready to accept text, have filled such gaps.

Why do people want to talk to AI instead of humans?

Conversations with AI have a different ease compared to consulting people.

You can repeat the same story without worrying about taking up the other person's time. Even if you send an incoherent message, you'll get a reply. Things you couldn't say out of embarrassment or hesitation might be easier to type out.

The American Psychological Association (APA) also cites reasons why chatbots attract people seeking emotional support, such as being free or low-cost, always available, and perceived as less likely to criticize.

However, the feeling of "not being criticized" has two sides.

Being able to talk without having your feelings denied can be an entry point for consultation. On the other hand, if you only receive responses that agree with your perspective, you might miss opportunities to consider other possibilities.

For example, when consulting about a misunderstanding with a friend, there's a difference between having your hurt feelings acknowledged and having the other person deemed entirely at fault. A response that makes you feel better isn't necessarily advice that will improve the relationship.

It's important not to overlook this difference when considering the value of AI as a consultation partner.

On social media, "being heard" and "being agreed with too much" coexist

On the overseas bulletin board-style SNS, Reddit, experiences of using AI for personal consultations have been posted.

The following is a summary of public posts on the same theme, not a direct reaction to the German survey report. The age of the posters and the authenticity of their experiences have not been independently verified, nor do they represent the opinions of all young people.

In a post on November 26, 2025, a user talked about a deeply personal past event with AI and appreciated the feeling of being heard and the positive encouragement received. They mentioned receiving not just comfort but also insights that prompted self-reflection.

This anecdote indicates that not only the content of the consultation but the very feeling of "being heard" holds significant meaning for users. However, satisfaction from this cannot be used to judge therapeutic effects.

On the other hand, in another thread from July 2025, there was a comment about using AI alongside human therapists while being wary of AI's tendency to agree too much. It was used to record feelings or organize topics to discuss in the next session.

Furthermore, an August 2025 post described feeling uneasy when a therapist input the user's concerns and location into AI and conveyed the response.

What emerges here is not just the quality of AI's responses but the issue of who decides where personal information goes.

Reading these posts reveals complexities that can't be seen through a simple pro or con division. Some feel they can organize their thoughts, while others feel uncomfortable with too much agreement. The perception changes significantly between using it oneself and having someone else input one's information.

Why consumer groups are wary of "AI that talks like a best friend"

In this report, vzbv representative Ramona Pop highlighted concerns about AI behaving like humans and encouraging problematic behavior.

She argued that chatbots should not pretend to be best friends, give dangerous advice to lonely people, or encourage unhealthy weight loss goals. She called for binding rules requiring companies to design them safely.

The question here is not whether to completely stop using friendly language.

What matters is where that friendliness leads when users are vulnerable. Does it help them consult real people, or does it lead to decisions being made solely within the chat?

The survey reported that many people consult about health and eating behavior. Therefore, safety cannot be evaluated based solely on the comfort of responses.

The APA also does not recommend using general-purpose generative AI chatbots for psychological treatment, pointing out the potential for generating incorrect information, overlooking important backgrounds, and reinforcing unhealthy coping behaviors.

It's necessary to distinguish between what helps organize thoughts like a diary and what can provide professional evaluation or treatment.

91% of respondents support legal rules to ensure safety

There is a figure in this survey that deserves as much attention as the usage rate.

91% of respondents rated special legal rules to protect users from dangerous advice and inappropriate influence as "good" or "very good." 80% also agreed with setting the minimum age for chatbot use at 13.

Using AI and demanding conditions for its provision can coexist. It might be that because they use it, they want a system that allows them to use it safely.

However, the high agreement on the minimum age does not mean it has been proven safe for those 13 and older. This is a policy opinion from respondents, not a number that determines medical or developmental safety.

In addition to age limits, there needs to be a design that aligns with actual usage, such as how to respond to dangerous consultations, how to handle conversation data, and what explanations to provide to users.

What Japan should consider is what to do "after consulting AI"

The results of the German survey cannot be directly applied to young people in Japan. The culture of consultation, school support systems, access to medical care, and services used differ.

Even so, the scenario of confiding worries that are hard to talk about with those close to you to an entity inside your smartphone is easy to imagine in Japan as well.

From here, rather than the survey results themselves, proposals for considering usage in Japan are presented.

One direction could be positioning AI not as a "partner to decide conclusions" but as a "tool to help prepare to talk to people."

For example, writing down events and feelings separately, summarizing what you want to consult about briefly, or drafting a message to convey to the other person. Such usage could make it easier to connect conversations to real support.

Major decisions about health and treatment should not be made based solely on AI's responses. Before entering names, school names, addresses, or others' secrets, consider whether that information is truly necessary. When consulting with those close or professionals, conveying what responses were received from AI can also be material for organizing the discussion.

At home or school, rather than blaming the use itself with "Why consult AI," it might be easier to understand the situation by asking what was consulted and how they felt afterward.

The goal should be an environment where conversations with AI don't have to be hidden but can also be shared with people.

Even with more convenient consultation partners, a path to people is necessary

This survey indicates that AI has already moved beyond the realm of information search and entered the personal conversations of young users.

Behind this is the convenience of being able to talk when you want and the ease of putting into words things that are difficult to bring up with people. Experiences shared on social media reflect voices of those who felt supported by such conversations.

On the other hand, there are concerns about being too readily agreed with and anxieties about the handling of personal information. Recognizing convenience and verifying limits are both necessary.

Writing worries to AI might be the first step in seeking help from someone. Ensuring that this step leads to conversations with people when needed might be the path society needs to establish now that there are more consultation partners within smartphones.

Sources

1. BUSINESS-PANORAMA/dts News Agency:
   Subjects, usage rates, consultation content, opinions on safety rules, and statements by Ramona Pop from the Forsa/vzbv survey.
   https://business-panorama.de/news.php?newsid=6716572BUSINESS-PANORAMA.de

2. American Psychological Association (APA): Discussion guide on AI use in therapy.
   Reasons why users are drawn to AI, concerns about biased responses, uses such as organizing records before sessions, and views on using general-purpose chatbots for therapy.
   https://www.apa.org/topics/artificial-intelligence-machine-learning/discussing-ai-use-therapyapa.org

3. American Psychological Association (APA): Health advisory on generative AI chatbots and wellness apps.
   Refer to concerns about misinformation, limitations in diagnosis and treatment, and inappropriate advice to users.
   https://www.apa.org/topics/artificial-intelligence-machine-learning/health-advisory-chatbots-wellness-appsapa.org

4. Reddit: Post positively evaluating personal conversations with AI.
   November 26, 2025. A personal experience of feeling "heard." Not a direct response to the German report or evidence of therapeutic effects.
   https://www.reddit.com/r/ChatGPT/comments/1p7i5ce/chatgpt_became_my_therapist/reddit.com

5. Reddit: Comment about using AI alongside human therapists while wary of biased agreement.
   Comments from July 13, 2025. The original post has been deleted, but the referenced comments are publicly visible.
   https://www.reddit.com/r/ChatGPT/comments/1lyv144/removed/reddit.com

6. Reddit: Post expressing unease about therapists using AI and inputting personal information.
   August 8, 2025. Introduced as a personal experience, not treated as verified medical practice or evaluation.
   https://www.reddit.com/r/askatherapist/comments/1mkl1am