What Teens Say About AI Companionship: Key Findings from Youth Co-Design Workshops in India and Singapore

Key Findings

• — 40% fear that depending too much on AI could make them stop thinking for themselves; loss of critical thinking was the top concern about AI's future, ahead of productivity (31%) and job displacement (27%).
• — 59% identified the absence of judgement as why someone their age might confide in AI, ahead of supportiveness (23%) and privacy or confidentiality (19%).
• — Asked why people form attachments, students ranked emotional language ahead of loneliness (32% against 23%).
• — On engagement mechanics the top three named were emotional language (26%), sycophancy (22%) and designed engagement (22%).
• — 69% expect privacy education to continue, and the framing of privacy concern was loss of control over one's own narrative rather than data collection.
• — 131 students aged 12 to 18 across 9 schools in 2 countries, 59% aged 15 to 16, 55% female and 42% male.

Methodology Notes

Workshops of 90 minutes each, run December 2025 to February 2026. The report states its own limits plainly: participants were not selected for representativeness, it is qualitative insight-gathering rather than a population study, findings may not generalise, and confirmation and social-desirability bias may be present because Cyberlite staff facilitated the sessions. Corporate provenance (Microsoft), with IMDA's Digital for Life movement named. The cover states only 2026; published_date is set to 2026-04-01 with month precision, the PDF creation date being 2026-04-10. The 29-page PDF (1,476,686 bytes) was downloaded and the figures verified in the text.

Sources