Issues in the Design and Implementation of Chatbots for Oral Language Assessment

Oral assessment in computer-assisted language learning is one of the best-known challenges at a technical and implementation level in an official language certification. The case of Spain is especially critical since the government has delayed for years the completion of the listening comprehension and oral expression test in the University Access Test (EVAU). This article presents first the evolution of oral tests at a general level, then a SWOT analysis of the potential of such implementation, and, third, how to implement them and the paper concludes that there is evidence that chatbots adapted to language learning can also be used for evaluation. Chatbot-assisted language learning with artificial intelligence adapted to voice recognition and its processing to obtain semi-automatic assessment supervised by the teacher can become a tool to be implemented within the language learning processes and / or included in the language certification methodology. Another interesting aspect could be the development and design of the interfaces of future adapted chatbots that must consider multimodality in the interaction between man-machine so that communication is effective and can be validated based on the knowledge available to the student.