For a multi-modal human-computer interface to better serve its user, it is first necessary to get a better understanding of the implicit aspects of human learning and memory. To clarify this claim, this paper first overviews relevant progress in cognitive psychology and presents our recent studies on the implicit learning of complex visual regularity in faces and cross-modal context effects on face and voice recognition. These examples are expected to illustrate the workings and power of an implicit learning mechanism and the intricate interaction of implicit and explicit components of face and voice memory. Implications of these results are then discussed for the design of human-computer interaction.