Speech perception and production may be linked. In this pilot study, we aim to investigate the nature of this link by examining how perception and production of VOT adjust in social interaction. In an online experiment with 2x2 conditions participants were asked to categorize nine stimuli on a VOT continuum between /tin/ and /din/ and produce the words din and tin before and after playing a game with a bot. During the game participants were trained on a sound to word correspondence with a bias either towards /din/ or /tin/. In two conditions participants alternated categorizing and producing the stimuli, while in the other two participants only categorized the stimuli. Significant differences in categorization between conditions occurred only during the game. Whereas significant changes in VOT can be observed between pre and post-test for three conditions. This could mean that perception and production do not adjust symmetrically over the course of an interaction.