Feel strongly about these letters, or any other aspects of the news? Share your views by emailing us your Letter to the Editor at [email protected] or filling in this Google form. Submissions should not exceed 400 words.

Midnight. A teenager in Hong Kong lies in bed, his mind racing with thoughts that won’t stop. What if I’m not good enough? What if everyone thinks I’m fine but I’m not? The worries feel too embarrassing to tell others and too heavy to carry alone. So he picks up his phone, opens the chatbot app and starts typing.

Generative AI is rapidly becoming many people’s first port of call for emotional support. In Hong Kong, a Chinese University of Hong Kong survey this year found record-high levels of depression and anxiety, with 55 per cent of respondents saying they would not seek professional help or were unsure. AI assistance is now the sixth most common option among those seeking help.

As a clinical psychologist, I see this pattern every week across young people in general. Generative AI chatbots give many of them their first real experience of being heard when they feel overwhelmed. They offer personalised analysis, validation and concrete suggestions in just a few exchanges. This is a genuine and valuable development in a city where emotional disclosure still carries heavy stigma.

Yet convenience is not the same as competence. Feeling calmer after a chatbot exchange is not the same as being properly assessed or safely supported. Research on chatbot sycophancy shows that endlessly agreeable systems can weaken reflection over time. The boundary is clear: AI may help with emotional first aid, but it should not replace professional assessment or ongoing support.

Professional bodies have been right to issue clear cautions. Most clinicians advise young people to treat these tools with caution and to seek human support when distress deepens. While caution is necessary, we have been slower to provide practical, accessible guidance on how to use these tools responsibly.