Most of us have realized at some point in our life that in the process of teaching, we are learning tremendously. Most of us are familiar with it. But, even though there are multiple online tutors, there isn't a single online artificial student to whom you can teach and in the process, learn!

Now the person whom you can teach may or may not be available and he/ she may or may not want to learn at that point. To overcome these challenges of flexibility, dependability and ability to scale, we needed to create an artificial student - Eklavya!

is an AI based chatbot which has an intellect of a sixth grade student who is willing to learn the topic of percentage. In the process of teaching, the real world student would go through a process of experiential learning which would improve the understanding level of the student.

Is teaching an Artificial Intelligence based chatbot more effective in learning and revision than other methodologies like solving homework problems?

Also, to make the research question more focused, we decided to focus on a single question: 'What is 75 percent of 3000?' The task was to ask a group of student to teach the problem to the chatbot and the remaining group to solve the homework problems. The effectiveness of both methods was compared.

Interview: Interviews of teachers, students and parents were conducted. The teachers and parents were interviewed which gave a deep dive into the teaching process, the struggles of learning that a sixth grader faces, the current teaching methodologies they use and the effectiveness of those.

The children were interviewed and they were told to teach. The innocence level kids have! :) They were also asked to share their experiences of learning the topic for the first time. What were the problems they faced? Did they enjoy doing the homework? What questions came to their mind as they were learning? If it helps if they are taught by someone of their age and so on.

Ethnographic studies: The data of all the interviews was recorded and analyzed. An important part of the ethnographic studies was to actually go into the classroom as the teacher taught the topic of percentage. I also visited a school for the underprivileged kids and the methods they use. (The teachers there need to be so much more creative than in their teaching process!)

Creating persona of the chatbot: Various literature was studied to get an idea into what would make the chatbot more interesting to the children. What are the specific traits that the chatbot should possess to make it interesting and fun. Here is the persona of the chatbot described briefly.

Eklavya is a character in Mahabharata (one of the two Sanskrit epics of Ancient India) who is considered to be the personification of learning, extreme levels of hard work and dedication. The persona of the chatbot is similar to Eklavya and based on his core qualities. Eklavya - the chatbot is a sixth grader who is 11 years old. He is eager to learn, humorous, talkative and loves to explore.

The principles of conversation were reviewed and analyzed. 'Fly on the wall' observation method was used for listening to the conversations between students and teachers. Along with it, the conversation analysis of observed and recorded conversations of the students was done.

The principles of human machine interaction were reviewed. It was fun to observe and understand the ways in which kids talk with the chatbots, the way they get excited and frustrated. It was an interesting process to understand how kids interact.

The chatbot was created using 'Google Dialogflow' and 'Google Actions'. An intellect of a sixth grade student who wants to understand the topic was replicated. The conversation design principles and all the previous observations were integrated to create the chatbot.

The video is of a real student of sixth standard chatting with Eklavya. The work is still in progress as the ways in which people teach division vary and it is a constant learning process for the code.

After having the students use the chatbot, it was observed that the students are very enthusiastic to teach the chatbot rather than solve homework problems.

The on field experiments demonstrated that having a conversation with the chatbot (with some guidance) is more effective in terms of revision compared to traditional methods of revision.