Computer programming a chatbot to improve social-communication skills in autistic children: A feasibility study

Abstract

Purpose A pilot study evaluated the feasibility of a curriculum that overtly teaches computer programming while covertly scaffolding social-communication skills for autistic children aged 8–12 years. Methods Participants were taught the Python programming language so they could program their own chatbots to greet a human user and discuss different topics, taking turns during the discussion, as though the chatbot were a human itself. The students were challenged with creating chatbots that pass the ‘Turing Test’, where a human evaluator would not be able to tell whether their chatbots were humans or computer programs. The curriculum included didactic instruction, peer-group discussion, homework and the chatbot project. Six autistic children participated in the six-session program. Feasibility was assessed using questionnaires and qualitative feedback. Results The curriculum is deemed feasible and desirable. There was no measurable change in social-communication skills immediately following the six-session program. Participants and their parents were highly interested in similar programs in the future, suggesting promising potential for further development and refinement. Conclusion A curriculum of programming a chatbot that also covertly scaffolds social communication is feasible for autistic children who are interested in computer programming.