The plausibility problem: Human teaching tactics in the 'hands' of a machine

This paper explores the Plausibility Problem identified by Lepper and his colleagues, namely the question as to whether teaching tactics and strategies which work well for expert human teachers can also work for machine teachers. We offer examples of the plausibility problem and argue that the issue arises in part from the impoverished expectations of students about what intelligent learning and teaching systems can achieve and in part from expectations about who should play what role in a learning interaction with a machine. We believe the problem will be reduced, though not necessarily eradicated, if systems are properly integrated within the overall educational context in which they are to be used, though further work is needed to establish this.