Benjamin Nguyen, Jie Yu, Mariam Khan,

Rezwana Kabita, Sanjidah Abdullah

Eng 21007 Writing for Engineering

Professor Maryam Alikhani

City College of New York

October 1, 2018

---

Abstract

The objective of this study was to build a chatbot that takes user inputs and responds to them in a personal way. We used Java in order to write the code and we tested our code in an online compiler. By integrating multiple If-Then statements into our code as our method, we took the user input and assigned them to different precoded cases to decide what the Chabot will respond with. Our results showed that with the multiple respond statements already coded in, the bot knows exactly how to respond and incorporates the input the user gives in to its responses. From this study’s conclusions, the chatbot we created is more of a foundation for larger scale projects that can create chatbots for lonely people as a form of therapy or for educational aid in the classroom.

 

Introduction

A chatbot is an artificial intelligence program that engages in conversation with users through text or audio. Chatbots are becoming increasingly common in modern day society through supermarkets, fast food chains, and even in our phones such as Apple’s Siri. These computer programs have revolutionized the marketing industry as it serves the purpose of saving the consumer’s time and money by instantly responding. For example, Domino’s Pizza’s Twitter allows customers to place an order of pizza by simply tweeting a pizza emoji (“The Complete Guide to Chatbots in 2018,” n.d.).

However, we designed an artificial interaction chatbot that focuses on the personal needs of the user regarding the user’s family and other desires, rather than through a marketing approach, as there are numerous marketing chatbots in use today. By providing a foundation of a simple chatbot, we can explore the future possibilities of more advanced chatbots that can focus on the mental health of users.

 

Materials and Methods

       In order to create our Artificial Interaction Chatbot, a certain software was used in order to create the Chatbot, but as with most computer programs, the most easily accessible was used out of convenience. Codeboard.io was used as the Integrated Development Environment in order to create the code of the project. Java was used as the primary coding language using a MacBook computer.

 

Procedure

       Two separate classes were created to run the Artificial Intelligence Chatbot. The Chatbot Runner class was used to run the code created in the Chatbot class in order to separate the backend from the front end of the code, making my code extremely user-friendly and neat. By creating two separate classes, the code can be organized and divided between running and computing of the actual code. The Chatbot Runner class also allows the person to add input and “speak” to the Chatbot.

As you run the Chatbot, the chatbot returns a greeting “Hello, let’s talk.” Then, different scenarios had to be coded in order to prepare response statements by the chatbot.

 

Enumerations of different scenarios coded for

If the user types nothing or no characters, the Chatbot responds with “Say something, please.”

If the user types a statement with the word “no” in it, the Chatbot responds with “Why so negative?”

If the user types a statement containing “brother”, “sister”, “mother”, and “father”, the Chatbot responds “Tell me about your family”.

If the user types a statement with the words “I want” + a single desire in it, the Chatbot responds with “Would you really be happy if you had” and then that desire in it.

If the user types a statement with the words “I” + an action or something else + “you”, the Chatbot responds with “Why do you ” + something + ” me?”.

If the user types a statement with the words “I want to  ” + an action or multiple desires, the Chatbot responds with “What would it mean to” and then the following actions and desires followed by a question mark.

If the user types a statement with the words “you” + an action or multiple desires + “me”, the Chatbot responds with “What makes you think that I ” + an action or multiple desires + ” you?”.

And lastly as a base case, if the user types a statement that does not belong to any of these scenarios, the Chatbot responds with one of four random statements, “Interesting, tell me more.”, “Hmmm.”, “Do you really think so?”, and “You don’t say.”.

These different scenarios that I have coded for ensuring that no matter what the user enters in, he/she will receive a response by the chatbot.

 

Results

• When the user types nothing or no characters, the Chatbot responds with, “Say something, please.”
• In case the user types a statement with the word “no” in it, the Chatbot will respond with, “Why so negative?”
• When the user types a statement with “father”, or “sister”, or any other title of a family member, the Chatbot responds with “Tell me about your family”.

• In case the user types a statement with the words, “I want cake”, the Chatbot asks, “Would you really be happy if you had cake?” and the question varies based on what comes after “I want”.
• When the user types a statement with the words “I love you”, the Chatbot asks, “Why do you love me?”
• When the user types a statement with the words, “I want to visit Canada,” the Chatbot asks, “What would it mean to visit Canada?”
• If the user types a statement with the words, “You like me,” the Chatbot responds with, “What makes you think that I like you?”
• When the user types a statement asking, “Will Apple do well in stock?”, the Chatbot responds with, “Interesting, Tell me more”.

 

Discussion

       While coding the Artificial Intelligence Chatbot, my group and I encountered several difficulties that dealt with the different scenarios that it had to be coded for.

One of the major problems was foreseeing what a person might say and taking the index of the main “subject” or “content” of the word and adding it to our prepared statements such as “What if you do if you had ________? ” Finding a creative approach in order to choose what index or location of keywords and predicting all the possible situations a person would ask of our chatbot was the most difficult. Scenarios such as what if the person types nothing or what if he/she says she wants something? The major challenge in coding for this project was determining the different scenarios to what a person would ask and coding for the appropriate response.

Other challenges we faced were trying to use the appropriate Java functions such as indexOf and using that to calculate the index of certain letters in a predicted statement by a user required the most mental math to calculate. Stepping outside of yourself and envisioning what an individual might say was the hardest part of the code, in order to write an appropriate response.

A significant benefit of this code is that the responses get more detailed and more accurate proportional to the more lines of code you add to the program.

 

Conclusion

Through the coding of a simple text-based AI chatbot, we get the basic idea of how chatbots are affecting our lives. Amazon’s Echo, Apple’s Siri, and metro card vending machines are a few examples of artificial intelligence chatbots that have successfully integrated into our daily lives and thus, have made our lives easier. People are continuing to design smarter AI chatbots that would create revolutionary changes in many other fields. We believe that AI chatbots can be developed in the medical and educational fields for the possible purpose of taking care of lonely people or even treating an autistic patient. However, such applications are currently impracticable because AI chatbots are still unable to understand human emotions or tones in conversation. AI chatbots function based on the rules the coders set. This is how our code functions in that it can only respond to certain statements with corresponding “answers” from the database we made. However, the same words can have very different meanings through various tones. If AI chatbots can be programmed to understand human emotion, it has the potential to help our society in unimaginable ways.

 

 

 

 

 

References

 

Ellis, R., Allen, T., & Tuson, A. (2007). Applications and Innovations in Intelligent Systems XIV.

London England: Springer-Verlag London Limited.

Fernandes, A. (2017, November 23). What is the future of chatbots? Retrieved from

https://www.quora.com/What-is-the-future-of-chatbots

Metz, C., & Collins. K. (2018, February 21). To Give A.I. The Gift of Gab, Silicon Valley Needs

to Offend You. Retrieved from

https://www.nytimes.com/interactive/2018/02/21/technology/conversational-bots.html

Nguyen, B. (2018). Chatbot Conversation [Digital Image].

Sprout Social. (n.d.). Topics. Retrieved from

https://sproutsocial.com/insights/topics/chatbots/#Sprout-s-Stance-on-Automation.

make a small caption for the screenshot and put someone’s name as a referenc