This article indicates that chatbots in interaction with an English second language teacher-evaluator offer for about 90% grammatically plausible, though often meaningless, absurd answers. Future performance improvements as “conversation practice machines” are forecasted.