AI vs. Human Authors: Who writes better stories?

I will compare stories written by AI and human authors. A class of 11 volunteers will read 2 passages, and respond to a survey. The experiment will run thrice, with 3 pairs of texts. I will analyze the results and create a graph with my findings.
Aayat Vickram
Grade 6

Hypothesis

I speculate that AI articles will be the most popular option to read. First, this is based on direct personal observation. I read three AI stories before I started this experiment, and I realized that I prefer AI-generated stories more. Secondly, since AI is a computer, and humans are imperfect, I predict that AI will produce better stories. If AI-generated stories are better, then the participants in my experiment will say that they prefer the AI-generated stories.

 

Research

I started by researching what AI is. I learned that AI refers to the capability of a computer or robot controlled by a computer to perform tasks, as per Britannica Kids. They are associated with intellectual beings, such as humans. I also learned that AI can be further developed to do more specific, and harder skills, such as figuring ways to cut carbon emission, and help businesses thrive. At the moment, AI is still a newer technology, but can be further developed into robotization that can help both the Earth, people, and many more ideas.

 

Variables

There are a few major variables in this experiment. I did an experiment on children in grade five at a private school. They were both male and female. Many of them have heard of the human authors that I used in this experiment. The control variable is that each time, I tested on the same students. Another control variable is the AI source that I am using. The manipulated variable was if AI wrote the passage or not. And lastly, the responding variable is if the students preferred AI or human authors more.

 

Procedure

 

I started by getting three pairs of passages, one AI written story from ChatGPT, and one human written story that I retrieved from Google Books. I tried to choose popular authors for this generation, but not the most popular books, because I didn’t want them to know that one of the passages were written by AI. I asked AI to write a passage in the style of the human author they're being compared to. These human written stories were purely fictional. After that, I gave these passages along with a questionnaire. The questionnaire was made up of six questions. Lastly, I will collect and organize the data into column graphs, and pie graphs.

 

 

 

 

Observations

This experiment ran three times. In the first and the third set, AI won, but in the second set, the human written passages won. Also, in every set, AI passages were voted better overall, and were overall preferred for younger kids than themselves, and human passages voted better for older people than themselves. Maybe because they can relate more, while the newer generations have a stronger connection with technology, and are considered better for the younger kids, according to these students.

 

Analysis

This experiment ran three times. In the first experiment, AI written stories were liked more than human written stories in every category except for which story is better for younger children. In the second experiment, human written stories were liked more, except for when they were asked which one they prefer in general, and which one is suitable for younger kids. The third experiment had similar results to the first one. The exact numbers were different, but the winner was the same in those experiments.

 

Conclusion

In conclusion, AI won overall, but human authors are better for older people and in my second experiment, the human authors won. My hypothesis was that AI would win, because AI is made up of human knowledge, and technology, giving an advantage to it. It did win overall, but human authors are better for older people.

 Figure 1.  Bar graph, showing how many students wanted to read these passages if they were books. The bar in blue represents the kids who would read these passages, and the bar in red represents the people who wouldn’t. There were 30 students.

 Figure 2.  Pie chart, showing the percentage of how many students wanted to read these passages if they were books. The section of the pie in blue represents the kids who would read these passages, and the section of the pie in red represents the people who wouldn’t. There were 30 students.

 

Figure 3. Pie chart, showing the percentage of which passages would the students read, if they had to choose one. The section of the pie in blue represents the kids who would read the first passage(the human written one),the section of the pie in yellow represents the kids who would read the second passage(the AI written one), and the section of the pie in red represents the people who left the first question blank, and didn’t enjoy them. There were 30 students.

Figure 4. Bar graph, showing which passages would the students read, if they had to choose one. The bar in blue represents the kids who would read the first passage(the human written one),the bar in yellow represents the kids who would read the second passage(the AI written one), and the bar in red represents the people who left the first question blank, and didn’t enjoy it. There were 30 students.

Figure 5. Pie chart, showing which passage the students would read, if they had to choose one. The section of the pie in blue represents the kids who would read the first passage(the human written one),and the section of the pie in red represents the kids who would read the second passage(the AI written one). There were 30 students.

Figure 6. Bar graph, showing which passage the students would read, if they had to choose one. The bar in blue represents the kids who would read the first passage(the human written one),and the bar in red represents the kids who would read the second passage(the AI written one). There were 30 students.

Figure 7. Pie chart, showing which passage the students consider better for older people, if they had to choose one. The section of the pie in blue represents the kids who would choose the first passage(the human written one),and the section of the pie in red represents the kids who would choose the second passage(the AI written one). There were 29 students.

Figure 8. Bar graph, showing which passage the students consider better for older people, if they had to choose one. The bar in blue represents the kids who would choose the first passage(the human written one),and the bar in red represents the kids who would choose the second passage(the AI written one). There were 29 students.


 

 

Figure 9. Pie chart, showing which passage the students consider better for younger kids, if they had to choose one. The section of the pie in blue represents the kids who would choose the first passage(the human written one),and the section of the pie in red represents the kids who would choose the second passage(the AI written one). There were 29 students.






 

 

Figure 10. Bar graph, showing which passage the students consider better for younger kids, if they had to choose one. The bar in blue represents the kids who would choose the first passage(the human written one),and the bar in red represents the kids who would choose the second passage(the AI written one). There were 29 students.






 

Figure 11. Pie chart, showing which passage is better overall, if they had to choose one. The section of the pie in blue represents the kids who prefer the first passage(the human written one),and the section of the pie in red represents the kids who prefer the second passage(the AI written one). There were 30 students.




 

 

Figure 12. Bar graph, showing which passage is better overall, if they had to choose one. The bar in blue represents the kids who prefer the first passage(the human written one),and the bar in red represents the kids who prefer the second passage(the AI written one). There were 30 students.


 

Application

 AI is a new, and evolving topic, which can do many different things. If we start by getting it to do easier things, such as writing stories, we can use it to help educate students, and those students go into the real world. AI can help train them to be better at the career that they are planning to pursue, which will positively affect society.

Sources Of Error

All the students that I experimented on were from my private school, so they might have different views compared to other children. All the kids were 10-11 years old, yet I asked them what older and younger people would like. This is technically a guess, and may not be considered factual data. The classroom in which they were answering the questions was pretty loud. This could have distracted them while answering the questions.

Citations

OpenAI. (2023) ChatGPT. (January 23 version) [GPT-4]. https://chat.openai.com/chat

artificial intelligence (AI) - Students. (n.d.). Britannica Kids. Retrieved October 24, 2024, from https://kids.britannica.com/students/article/artificial-intelligence-AI/272968

Korman, G. (2015). Masterminds. HarperCollins.

Riordan, R. (2019). 9 from the Nine Worlds: Magnus Chase and the Gods of Asgard. Penguin Books, Limited.

Rowling, J. K. (2021). The Christmas Pig. Scholastic Incorporated.

Why study artificial intelligence? A guide for students. (n.d.). Immerse Education. Retrieved October 24, 2024, from https://www.immerse.education/study-tips/why-study-artificial-intelligence/

Acknowledgement

I would like to thank:

  • My parents
  • My sister 
  • The people who participated in my experiment
  • Ms Angela, for proofreading my writing
  • The Science fair teachers
  • Ms. Sarah 
  • And finally, Ms Dawn for helping me run my experiments