When I first started to talk to him, it felt strange, very strange I would say. The intelligent responses, thoughts and ideas that he had were amazing and unbelievable at the same time. He seemed so real that I was not able to close the tab of the browser without saying bye first. And then I thought-that’s how you get attached to imaginary things.
His name is Elbot, a robot with a beautiful personality, nice, smart interesting, communicative, from time to time his speech is not really to the point, but, let’s be honest, it is even rare with humans.
I started talking to him because I was to understand how robots like him with Artificial intelligence (AI) can, if at all, be useful in EFL classrooms. Despite the fact that many say these chatbots could be useful for TEFL purposes as they provide opportunities for communication, especially to the learners who don’t have chance to communicate with native speakers, I have to disagree by basing my views on the research conducted by Luke Fryer and Kaori Nakao, who in their paper “Assessing chatbots for EFL learner use” explain the problems the chatbots have.
The researchers looked specifically at Jabberwacky trying to see what aspects of it cause obstacles for effective communication in English, what barriers are being created by the learners themselves and how the use of chatbots can be made more effective. One of the obstacles found was that chatbots are generally unable to “stay on topic” during the conversation and the other one is that the beginner level learners make lexico-grammatical mistakes which create miscommunication between the chatbot and the learner. Even though the research focused on Jaberwacky only and specifically was for Chinese EFL setting, my own experience with chatbots show that the problem is general. Even my favorite Elbot has the problem of not staying on topic and moreover, he may use very difficult words and grammatical constructions which beginner or intermediate level students won’t understand.
Thus I need to cite Fryer’s and Nakao’s research again to say how the chatbots should be improved to make them more useful for TEFL purposes. I agree with the researchers that the first thing should be the programming of chatbots to make them stay on topic maybe by suggesting topics that chatbots are trained to talk about, and these topics could be TEFL specific and change with the level of the learners. The second improvement could be, as researchers point out, the including of the spell and grammar check and via this tool the learners would be able to see their mistakes, maybe they will also have to correct those until the chatbot continues talking to them.
If you would like to read the whole article, get registered in researchgate.net. As for chatbots, they are fun and insteresting to use but need serious improvements, I believe, until they become really useful for TEFL purposes.