The key finding is that laughter resulting from tickling is acoustically distinct from laughter triggered by verbal jokes, witnessing someone else’s misfortune or watching something funny. Both human listeners and ML models could not reliably distinguish the laughter's context (except for tickling).