Implicit mind perception alters vigilance performance due to cognitive conflict processing

Knowing the internal states of others is essential to predicting behavior in social interactions and requires that the general characteristic of ‘having a mind’ is granted to our interaction partners. Mind perception is a highly automatic process and can potentially cause a cognitive conflict when interacting with agents whose mind status is ambiguous, such as artificial agents. We investigate whether mind perception negatively impacts performance on tasks involving artificial agents due to cognitive conflict processing caused by a potentially increased difficulty to categorize them as human versus non-human. Experiment 1 shows that an ambiguous humanoid stimulus negatively impacts performance on a vigilance task that is known to be sensitive to the drainage of cognitive resources. This negative effect on performance vanishes when participants are pre-exposed to the stimulus be-fore the vigilance task (Experiment 2 and 3). The effect of pre-exposure on performance recovery is independent of whether participants explicitly resolve the cognitive conflict by answering mind-related questions (Experiment 2) or implicitly by judging the stimuli on a set of physical features (Experiment 3). Together, the findings suggest that mind perception is so automatic that it cannot be suppressed even if it has negative effects on cognitive performance.