Article

Learning_with_Robots

Learning with Robots

Introduction

Working with robots is no longer the stuff of science fiction. The first industrial robot was designed by George Devol in 1954. More recent progress in artificial intelligence techniques has produced robots that can respond to changes in their environment instead of just repeating the same task over and over again on the factory floor.

Using robots for educational purposes is not new. The widely recognised Logo Turtles were described by Seymour Papert in 1980. The thinking behind the development of these was based on constructionism.

This theory of learning states that people construct their own understanding of the world. They do this through making objects that are tangible and shareable. They become active creators of knowledge who explore, question, and assess what they have learnt. Schoolchildren use Logo Turtles to solve problems. Doing this helps them to understand basic mathematical concepts.

Creating robots that perform well at specific tasks can be an exciting and collaborative learning activity. The RoboCup and RoboCup Junior competitions are now well established. Participants work in teams to produce robots that specialise in playing football or exploring a maze. RoboCup competitions also include events in which robots take part in an artistic or theatrical performance. These events recognize the growing potential for communication and expression that is possible with these technologies.

New practices

Robots have a long history in education, but the breadth of potential uses is now increasing rapidly. Artificial intelligence (AI) has been integrated into a variety of robots, creating an exciting space for learning together. For example, CHiP the Robot Dog has been developed to be a pet.

These robot dogs can share their thoughts, such as ‘I would like to go for a walk’, via an app. They develop their behaviour over time; respond based on how their owners interact with them, and can be trained.

Advances in design and hardware support natural forms of interaction or dialogue with robots, using speech, gestures, and emotional expressions. The SoftBank NAO robots can speak and understand 20 languages, using cameras to recognise people and objects. SoftBank states that learners perceive the robots as friendly and non-judgmental, which gives them confidence when answering questions. Robots designed as social communicators provide opportunities for language learning. The robot can act as a tutor, available at any time the learner wants to chat.

Robots are still a starting point for learning programming skills, but this can now be achieved in increasingly natural ways. For example, children can develop the understanding of a Photon robot by increasing its skills so it can complete challenges. The intention is that children learn by teaching the robot.

These latest developments extend possibilities for humans and machines to collaborate. In the field of medicine, robots can take on routine tasks, freeing humans to spend time with patients. The same is true in education. Robots can take on routine tasks, freeing teachers to spend more time with learners.

It’s now beginning to be possible, in some cases, for robots to take on the time-consuming job of skills assessment. Rather than simply demonstrating a task, learners can demonstrate their understanding by teaching the necessary skills to a robot.

This is possible due to new breakthroughs in Programming by Discussion. Robots can now learn not only from trial and error, but also from conversation with a human teacher.

This mimics the way human teachers behave with pupils where pupils understand how to go about solving a problem after being given verbal feedback on each of their attempts. This way of working has been well documented by Jerome Bruner and has been termed ‘scaffolding’.

Challenges

Advanced robotics equipment can be very expensive and can also require support to set up and maintain. This may mean that it isn’t possible to use particular technologies in many educational contexts. There are, however, many lower cost robots entering the mainstream market, and also many kit-based approaches aimed at supporting creativity by providing the building blocks for a range of robotics projects.

Many people have an inherent distrust of advanced technologies, and the combination of robotics and AI certainly can provoke strong reactions. The ‘uncanny valley’, in which certain levels of humanlike appearance in a robot can trigger negative responses in people, has long been recognised and debated.

This could limit the use of robots in education, but it is a problem that needs to be overcome in order to benefit from robots’ potential for promoting understanding. Learning with robots offers us opportunities to develop our judgement and ability to interact. It also offers a means to consider big questions related to the responsible use of artificial intelligence.

Citation:

Ferguson et al. Innovating Pedagogy 2019: Open University Innovation Report 7. Milton Keynes: The Open University.