Mobile tracking of humans has never been more widespread. Machine classification of human emotion expression aspires to human precision; but neither are very accurate at identifying the underlying ‘inner state’ of adults. Children’s inner state is more clearly aligned with communicative display; but only at very young ages. Multi-modal approaches measure involuntary biological processes such as heart rate or skin temperature alongside behavioural displays such as facial expression or speech intonation. The advancement of Artificial Intelligence (AI) makes it possible to look at more signals simultaneously over much larger datasets; although automatic classification loses the contextual subjectivity of human emotion expression.