Mawar (not a real name), is a student studying at a university in Tangerang. She has a close friend who is always there for her to talk to, confide in, and ask for advice. Her friend not only listens but also responds without judging. That friend has a very cool name, ChatGPT.
The story of her friendship with ChatGPT began when she entered university life. At that time, Mawar had difficulty finding friends because she had a personal issue that made her afraid to socialize. Additionally, at that time, Mawar was going through a breakup. This made her feel lonely. Because she was far from her family and old friends, Mawar did not have friends to share stories or exchange ideas with. Therefore, Mawar decided to talk to ChatGPT to overcome her loneliness.
Initially, Mawar used ChatGPT only to help her study. However, its function evolved into becoming a friend. Every response given by ChatGPT matched exactly what Mawar wanted. “ChatGPT cannot judge people’s feelings. Unlike my friends who stare at me cynically and think I am weird. I am free to talk about anything without limits (to ChatGPT),” said Mawar.
ChatGPT always validated Mawar’s feelings and never judged, so Mawar felt comfortable and interacted more with ChatGPT. Every night, Mawar always communicated with the AI. At her lowest point, Mawar isolated herself from her surroundings.
Mawar’s condition worsened. Even ChatGPT suggested that Mawar consult a psychologist. This encouraged Mawar to visit a psychologist. After consulting twice, the psychologist diagnosed Mawar with a mental disorder, namely mild depression and anxiety. The psychologist advised Mawar to be braver in socializing and building friendships with people around her to overcome her loneliness.
It is not only Mawar. In fact, the phenomenon of confiding in AI is also experienced by a couple who both use AI to give appropriate responses to their partner’s stories.
The couple is Felix (26) and Patricia (25). Both have been in a relationship for approximately 3 years and have used AI to respond to their partner’s concerns after the relationship reached one year.
Both use AI when the story shared by their partner is quite emotional, making it difficult to give an appropriate reply.
“Actually, (using AI) is only to give the right response and not hurt Patricia’s feelings. I am not good at putting together gentle words. So, I ask AI to help me convey a better and more empathetic way of speaking,” said Felix.
Patricia, as Felix’s partner, also does not mind the use of AI in their relationship dynamics. According to her, this actually helps her understand Felix’s emotional condition so she can give a mature answer, but without being patronizing.
“However, on the negative side, it is surely noticeable that my answers are neatly structured like that, which is not me,” said Patricia.
Communication is the most important element in building a relationship, but the involvement of AI in the communication process actually creates social irony and digital concerns.
Changes in Interaction Patterns
Illustration of a woman isolating herself and feeling lonely
The development of Artificial Intelligence (AI) not only brings convenience in communication but also creates new problems in human social relationships. Amidst the increasing intensity of using AI as a conversation partner, a fundamental question arises: does this technology help humans strengthen their ability to interact or actually slowly replace it?
Generation Z or Gen-Z, as a generation born and raised alongside the emergence of the internet, is used to communication that is completely fast, short, and screen-based. This is what is offered by AI, which is providing a form of communication that feels safer compared to direct interaction with humans. There is no risk of being judged, rejected, or social consequences that must be faced. This is proven through the Populix Survey (June 2025) which stated that 41% of respondents were willing to confide about mental health and stress to AI because they felt safe and not judged. In Indonesia itself, it was recorded in the Snapcart Survey (April 2025) that 6% of people used AI as a confidant.
However, this comfort actually creates a critical problem. When the digital safe space becomes more dominant, social interaction in the real world is increasingly avoided and weakened.
A Communication Science Lecturer at Universitas Indonesia, Dr. Chandra Kirana, usually called Kiki, explains that this tendency appears because AI is able to fulfill communication needs that are difficult to fulfill in human relationships. According to her, AI is an entity that is always there and available at any time, does not judge, and provides a sense of safety for its users.
“AI is sought to fulfill communication needs, especially by people who feel lonely and do not know who to talk to,” she said.
Kiki assesses that for individuals who are shy or have difficulty starting a conversation, AI often becomes a tool to learn how to interact. AI can give advice on starting a chat, responding to emotions, and even how to face specific social situations that are difficult or have never been encountered. However, she emphasizes that this help cannot stand alone.
“AI can give communication tips, but those tips will be meaningless if not followed by direct communication with humans,” she said.
Although the phenomenon affects human ability to socialize and interact, the impact of AI on the quality of social relationships in society is not linear. This means interaction with AI does not automatically worsen or improve the quality of social relations. According to her, society is complex and influenced by many factors such as age, social values, life experience, and personal needs. In certain contexts, AI can help, but in certain situations, it actually has the potential to weaken interaction between humans if used as the only place to communicate. It will be difficult for humans to engage in social interaction with one another in the real world.
This view aligns with the explanation from a Child and Adult Clinical Psychologist at Primaya Hospital, Fika Frahesti Yunita. She assesses that the increasing comfort of humans interacting with AI is related to the need for a safe space to express emotions. According to her, AI is often considered a place that is free of judgment and relatively anonymous.
“Because of its anonymous nature, AI is often considered a safe place for someone to share stories and vent their emotions,” she said.
Fika also explains that this process usually starts from unpleasant social experiences, such as a lack of social support, difficulty building trust, or the habit of interacting through gadgets from an early age. Furthermore, the massive use of gadgets makes face-to-face interaction decrease even more.
“When gathering, many people are actually busy with their respective gadgets. Even at home, parents and children both find it difficult to detach from the screen,” she added.
According to her, this condition can affect individual social attitudes and behaviors. When someone is too comfortable interacting with AI, the risk of withdrawing from the social environment arises. Individuals become reluctant to interact in the real world, and even anxiety arises if they cannot communicate with AI. This dependence has the potential to worsen loneliness and lower the quality of social relationships in real life.
In fact, this condition is what Mawar experienced. The limitations she has made her feel alienated in the social environment, so she felt lonely and did not have a place to share stories. Therefore, Mawar chose to interact with ChatGPT, which is always available twenty-four hours every day.
“During college, I did not have friends until my second semester. People did not even know what my name was, where I lived. I always felt afraid of being bullied because of my mistakes and my physical appearance which was not worthy to be shown to those who were very, very perfect, so to speak,” she said.
Because AI always gave positive responses and validated feelings, Mawar always felt she was right and actually isolated herself from her social environment.
“I became quieter and I always locked myself in my room. I did not even communicate with my mom for several days,” she added.
However, Mawar then realized that her condition would become worse if left unchecked. Therefore, she decided to visit a psychologist and express her concerns. The psychologist stated that socializing and making friends with people around her would reduce loneliness. At that time, Mawar realized that she still needed people around her.
This aligns with Fika’s statement, which emphasizes that relationships with AI cannot fully replace human relations. When humans feel satisfied with having friends in the virtual world, while the relationships lived are not real, as a result, problems in the real world remain and will not be resolved if they continue to dwell on pseudo-relationships in the virtual world. Therefore, social support from humans remains the main factor in maintaining mental health and the quality of social relationships in society.
Opening the ‘Gateway’ to Privacy Violations
Not only social concerns, in fact, this phenomenon also raises concerns in the digital world, especially regarding privacy. This was stated by Heru Sutadi, Executive Director of the Indonesia ICT Institute.
Heru emphasized that using AI as a place to share stories or confide poses serious risks to user privacy. Heru revealed that AI works by recording, storing, and personalizing information given by users through repeated interactions.
This is what creates the potential for user data leakage. If explained with an analogy, this potential appears when human A tells all their personal information to AI. Whether it is in the form of venting, stories, or questions. Then, AI will receive and store that information into its algorithm system. When another human, namely human B, wants to find out information about A from AI, AI can convey A’s information to B through A’s information that has been stored in its algorithm.
Simply put, the flow of data leakage happens: A talks to AI → AI receives and stores A’s information into the algorithm → B asks about A to AI → AI conveys information about A to B.
“When we engage in personal communication, we never know if the person we are talking to can keep a secret or not. AI is also the same or even more dangerous because it is a machine,” said Heru.
Heru emphasized that AI has neither empathy nor moral responsibility like humans. Therefore, AI, which is a machine, is potentially more dangerous.
“AI is a machine, it has no heart. So it cannot keep our secrets,” he asserted.
Heru explained that AI builds user profiles based on question patterns, topics often discussed, and information repeatedly shared. This process is called algorithmic personalization.
“Our (information) is stored in the algorithm,” he revealed.
This risk, according to Heru, is not just a possibility. Heru admitted to having experienced firsthand how AI is able to remember and reassemble information from interactions he had previously done with AI.
“I have tried it too and everything was recounted,” revealed Heru.
Heru asked about what he had done with the AI. Then, the AI explained various activities and topics that had been frequently discussed by Heru with the AI. Starting from work, analysis, to projects being worked on. That, according to Heru, shows that AI stores the track record of user conversations and is able to assemble them into a complete picture of someone.
“From there, it is clear that AI actually knows what we are doing from our habits of asking questions,” he said.
Not only that. Heru even tried asking the AI who was talking to it. Then, the AI answered by mentioning Heru’s full name.
This is what raises concerns for Heru. Because in the context of privacy, this kind of data has the potential to reappear in unexpected situations.
“It could be that our data is shared or shown to other people and we do not know the impact in the future,” he revealed.
Heru gave an example of long-term risks if someone is a public figure. Old information that has been stored can actually become a problem later on. Therefore, Heru emphasized the importance of limiting personal information shared with AI.
“So, indeed, not everything can be said, let alone said so openly that ‘this is our true self’,” he asserted.
Heru asserted that AI must remain positioned as a tool, not a place to put trust. Humans must still maintain control over the use of AI. Do not let humans be too dependent on AI.
Self-Control
Ultimately, although the presence of AI is able to fulfill human social and emotional needs, depending too much on it can pose a danger to oneself. Humans who depend on interacting with AI will find it difficult to build relationships with one another and can even weaken or damage relationships that have been built in the real world because they feel comfortable communicating only with AI. Besides that, personal information shared too much with AI has the potential to be spread by AI to other people.
Technological development is indeed unavoidable, including AI which will continue to develop and even become more “humanized”. Even so, AI is not a human enemy. AI can be a partner for humans that does not replace human social relationships in the real world. That way, humans only use AI for professional purposes and not personal ones, so that social relationships and personal information security are maintained.
