Social robots, ethics and students’ data

Social robots talk, look at the student and react. In 2018, the review by Belpaeme and colleagues showed positive effects on targeted learning, such as vocabulary, sometimes close to those of a human tutor on narrow tasks. It also noted their technical limitations. In 2020, a review by van Straten and colleagues showed that children easily form a relationship with a robot. They feel close to it and trust it.

This relationship raises ethical questions. In 2016, Amanda Sharkey examined the idea of robot teachers. She identified risks to privacy, the child’s attachment to the machine, a form of deception when the child believes the robot cares about them, and the loss of human contact. She judged the robot as a tool serving the teacher to be defensible, but not the robot that replaces the teacher.

The law already governs some uses. In the European Union, the AI Act has prohibited, since February 2025, the use of AI to recognise people’s emotions in educational institutions. An AI that assesses learning or guides a student’s path is a high-risk system. Since Regulation (EU) 2026/1744, these obligations apply from 2 December 2027. A Swiss school can be concerned when it serves students in the Union.

In Switzerland, a public school applies the cantonal data protection law. The federal law, in force in its revised version since 1 September 2023, applies to private schools and to the companies that supply the robot or the online service. In both cases, a student’s voice, image and activity traces are personal data.

• Does it work offline? If not, what data does it send, and where is it processed?
• Does it record the students’ voices, images or location? How long does the supplier keep them?
• Does each student need an account, and with what information?
• Has your school or your canton authorised it?
• Do the parents know what is recorded?