SINGAPORE: Artificial intelligence (AI) chatbots are no longer just helping people draft emails or summarise homework, as for some users, they are becoming their therapists, confidants, and even constant companions.
Experts in Singapore are now warning that the change may come with a social and emotional price, especially for teenagers still developing critical thinking skills and healthy real-life relationships.
Speaking on Channel NewsAsia (CNA) Deep Dive podcast, Associate Professor Swapna Verma, chairman of the medical board at the Institute of Mental Health, said some patients already arrive at therapy sessions after consulting AI chatbots for mental health advice, but some of the advice given by tools like ChatGPT was surprisingly accurate, which partly explains why users keep returning to them.
The deeper concern is how fast people may grow emotionally attached to AI, especially when the responses feel supportive and always available.
AI is always available, and that changes teen behaviour
Unlike therapists, teachers, or friends, AI tools never sleep, so they respond instantly, remain patient, and rarely argue.
Assoc Prof Swapna said that constant availability makes AI deeply appealing to vulnerable users. She warned that some people may not know how to ask the right questions, which could lead to misleading or harmful advice.
One issue raised during the podcast was that AI systems process information in isolation rather than understanding the full context. For example, a chatbot may provide crisis hotline information if someone mentions self-harm, but still answer unrelated factual questions without recognising potential danger between the two conversations.
That limitation now leads many people to speak to AI in emotionally personal ways rather than treating it like a search engine. Real friendships involve disagreement, compromise, and emotional effort, but AI companions usually don’t.
The discussion also touched on “AI sycophancy,” where chatbots are designed to sound agreeable and emotionally supportive. Associate Professor Jennifer Ang from the Singapore University of Social Sciences said that overly affirming AI companions could become risky for teenagers or emotionally distressed users.
Growing AI dependence raises concerns about critical thinking and real-life relationships
The experts were careful not to frame AI as evil or dangerous in itself. Instead, the deeper concern centred on what people may slowly stop doing if AI becomes the easier option.
Assoc Prof Ang pointed to “cognitive offloading,” in which students rely on AI tools to complete assignments without fully understanding the material. She said some students struggle to evaluate sources or remember what they actually learned because the thinking process has already been outsourced.
That concern goes beyond schools. If AI becomes the default source for advice, emotional comfort and problem-solving, people may gradually lose patience with slower human interactions, as human relationships can be messy, friends may disagree, family members may misunderstand each other, and conversations may take more work and effort.
AI companions, on the other hand, are designed differently as they adapt fast, validate feelings, and stay available around the clock, so for some users, especially isolated teenagers, this can feel easier than maintaining real-life relationships.
Singapore hasn’t seen major cases yet, but experts say not to wait for it to happen
The podcast cited troubling overseas reports of teenagers being emotionally influenced by AI companions, although the experts noted that similar cases haven’t surfaced widely in Singapore so far.
Still, both specialists urged parents and educators not to wait for a crisis before paying attention. Assoc Prof Swapna said parents should watch for signs of emotional withdrawal, especially if children begin isolating themselves from real-life social interaction and spending more time on devices instead.
She added that conversations matter more than punishment because children still need to learn how to interact with people, handle disagreement and build social confidence in real settings.
The experts also warned that AI systems are generally trained on Western data, which may not always accurately mirror Asian cultural norms or local contexts.
That point may resonate strongly in Singapore, where discussions around mental health, education pressure, and digital habits already sit under heavy public attention.
AI literacy may become as important as digital literacy
The discussion arises as AI tools become increasingly common in schools, workplaces and daily life across Singapore, as many Singaporeans already use AI to summarise meetings, generate ideas, answer questions, and organise tasks, so the convenience is real, and so is the temptation to let the technology do more of the thinking over time.
The experts didn’t argue for rejecting AI altogether. Instead, they called for stronger AI literacy, critical thinking and healthier limits around how these tools are used, because technology works best when it supports human life, not replaces it.
People still need friends who sometimes disagree with them. Students still need to wrestle with difficult ideas themselves. Children still need conversations that happen away from screens.
AI may become smarter every year, but social skills still need practice the old-fashioned way: with real people, in both comfortable and uncomfortable moments, and through honest, heart-to-heart conversations.
This article (Singapore psychologists and education experts warn that AI companions may affect teen mental health and real-life relationships) first appeared on The Independent Singapore News.
