4.3.2 (HL)—Technology and development

Technology can influence development of self and attachment by changing feedback, social comparison, caregiver contact and opportunities for exploration. Artificial-intelligence models can simulate or classify developmental patterns, but a computational fit is a test of specified predictions rather than a copy of a developing child.

A platform may supply identity communities and responsive contact, or expose a child to quantified approval, exclusion and surveillance. Video contact can maintain a relationship across distance but does not reproduce touch, shared routines or caregiver availability. Effects depend on age, activity, relationship quality, access and offline context.

Researchers might train an AI model on children's task responses to compare gradual learning with stage-like transitions. Success on held-out data supports predictive value under those conditions; biased training samples, opaque features or shortcut learning can make the model inaccurate for other cultures or tasks and cannot establish the psychological mechanism by itself.

Screen time is not a developmental mechanism, AI accuracy is not proof of human-like cognition, and online contact is not automatically secure attachment. Define the activity and outcome, compare plausible pathways, protect child data and avoid replacing developmental or relationship evidence with a technology label.