The ethical concerns with robot teachers in primary schools are: privacy, attachment, deception, inappropriate dealings with humans, control, and accountability. Four roles for robot teachers are: classroom teacher, companion and peer, care-eliciting companion, and remote teacher. The paper concludes that robots can impact a child’s privacy, particularly if sophisticated sensors and thorough data collection/storage are employed. The paper also concludes that students could become emotionally attached to a robot acting as a companion, to a robot they care for, or even to a robot acting as a teacher.

Almost by definition, a robot teacher deceives the pupil. An ethical issue is how a deceived pupil’s ability to function with humans is affected; the pupil may, for example, expect people to always forgive. Having a robot control people is controversial and yet being a teacher implies having some control. One could argue that a robot might be fairer than a human, never angry or vindictive, but fair decisions seem to imply empathy and we are far from developing an empathetic robot. Legal accountability for a robot’s actions, even for a self-driving car, is not at all clear. Among possible reasons stated for using a robot teacher, only telepresence, a robot controlled electronically by a distant operator, seems justifiable. An example is teaching English to children in a remote classroom. Another use that may have promise is having pupils teach a robot.

The paper reviews existing work clearly, makes valid arguments, and in general is a pleasure to read.