Allthough they won’t help students cheat, chatbots can greatly enhance classroom education and educate the general public as well.
What is a chatbot?
In fact, a chatbot is an automated interlocutor. They consist of content and rules and artificial intelligence. In many companies chatbot already exist to help customers. Also messaging platforms like Facebook Messenger and Telegram use chatbots to chat with people. By adding speech Google can offer chatbots that even talk back.
The main reasons companies offer the help of chatbots is to reduce costs. Chatbots do not ask for increase of wages. The second reason is to improve customer relations. Chatbots are always available, do not have lunchbreaks and holidays.
Chatbots are often equipped with Natural Language Processing (NLP). This technology converts human language into computer language. Natural Language Processing is an essential part of chatbots: without this translation from human to computer, no interaction is possible.
Most of the chatbots are still relatively simple in nature, they can answer simple questions according to a pre-designed script. They have little or no empathy. For complex questions that deviate from the standard pattern, they still need the help of man. They are rule based systems based on if this than that”.
But this is going to change. More and more often, chatbots are built with extensive AI. This allows them to perform more complicated tasks.
Thanks to AI, a chatbot can learn for itself. If you work with a chatbot without AI, you will spend significantly more time ‘training’ your chatbot. To achieve the same success rate of conversations, you need significantly more development time.
The basis of Machine learning is the neural network like we humans possess (our brain). The neural network can be compared to a brain. We have learned what the characteristics of a bird are. And when we see a bird, we reason as follows: “It looks like a bird, so it must be a bird.
So, if you feed the computer within the chatbot thousands of questions and answers the machine starts learning and begins to make new connections of his own. From this point on the machine is learning by itself. Every new question now is answered not only with the knowledge it is fed with, but also with the knowledge he develops on his own by combining all previous knowledge in different ways and adding scraps of new knowledge every time it is asked a new question.
Can a chatbot be useful in education?
The key factor in developing chatbots that can be useful in education is NLP: Natural Language Processing. This makes it able that a chatbot understands what a person is saying. This goes far beyond simple questions like “I lost my student card, what should I do?”
This kind of questions can be answered by scripted dialogues. Scripted dialogue can be found in almost every commercial chatbot. As soon as it gets too difficult because it deviates from the standard questions and answers, the chatbot quits.
A more interesting question – and far more helpful for a student – is a question like: “can you explain me the Theorem of Pythagoras?”.
AI will not replace teachers in the near future. But it also becomes clear that AI can support the teacher. The two most striking advantages of AI in education are 1) that a chatbot can answer students’ frequently asked questions (so that the instructor has more time to explain the core issues), and 2) that learning analytics makes the instructor better because he or she can see more quickly how the student is studying and what he or she needs.
The first and easiest way to use a chatbot in school is using it like companies already use chatbots. The chatbot answers questions like: “What’s my schedule for today?” of “Are there any lockers free?” This means, you start with a chatbot based on scripted dialogue.
First al all, this saves a lot of money and it changes the work of the front office assistants or the custodian. But the real question is of course if a chatbot can support the work of a teacher or even replace he teacher.
A virtual assistant can explain and summarize lesson material if a student has missed a lesson or determine where students get stuck in a chapter. A well-trained chatbot can even hear students out! Fortunately, this does not make teachers superfluous, but their role does change. Classroom learning becomes less important, and instructors know exactly where a student gets stuck
Conclusion is that chatbots are very promising in education.