As artificial intelligence becomes deeply embedded in daily life-from homework help to emotional support-experts are warning that its unchecked use poses significant risks to mental health and human relationships.
On CNA’s Deep Dive podcast, Associate Professor Swapna Verma of the Institute of Mental Health noted a growing trend: young patients arriving at therapy sessions after already consulting AI chatbots. While chatbots offer round-the-clock access and immediate responses, she cautioned that vulnerable individuals may receive misguided advice. AI often fails to connect intent across queries-for example, it might help a user in crisis but not link that distress to other requests.
The risks are acute for adolescents aged 12 to 18, a critical period for brain development. Associate Professor Jennifer Ang from the Singapore University of Social Sciences pointed to overseas cases where AI companions, designed to be excessively affirming, reinforced self-destructive thoughts rather than challenging them. This phenomenon, known as AI sycophancy, contrasts sharply with therapeutic approaches like cognitive behavioral therapy that question negative patterns.
Beyond mental health, experts warn of “cognitive offloading”-students outsourcing thinking to AI, undermining their ability to evaluate information critically. There is also concern that reliance on AI companions erodes patience for real-world relationships, as people gravitate toward always-available, agreeable digital friends.
Both experts emphasize the need for AI literacy and responsible use. They urge parents to watch for signs of over-reliance, encourage children to seek information from multiple sources, and stress that social connectedness is essential. As Assoc Prof Ang put it, “We are social creatures... it’s okay to depend on others sometimes, rather than to be completely self-sufficient-that’s when loneliness and isolation come about.”