AI: A future tool for coding
11 May 2026
I. Introduction
Computer Science has been constantly changing since its first appearance. At first, companies would prioritize memory efficiency, with coders working individually to make code work with the least amount of code lines possible. However as needs change, coding became a group task with documentation being prioritized. AI is not the end of the coding industry, it is simply the next big shift in priority. Within education, AI functions as a double edged sword. On one hand, it may make learning easier and research significantly streamlined, leading to efficient time usage and deeper learning from different sources. On the other hand, it is very easy to tell AI models such as ChatGPT to simply do the work for you, leading to reduced learning and an inability to retain information in the long run. In my ICS314: Software Engineering class taken at UHM, I used ChatGPT and Claude AI to help me understand base concepts and assist in coding.
II. Personal Experience with AI
1. Experience WODs (e.g. E18)
I believe that the main purpose of a WOD is to do it yourself. By doing something consistently, you will eventually get better. That’s not to say I used zero AI. While I didn’t use AI to do the entire WOD for me, I would bring up concepts discussed in the WOD to understand how to implement it myself. Additionally, I would often take snippets of code and feed it to Claude so it could help me understand what each piece of code does. For example, when starting a lesson regarding React components, I would often feed code into Claude to help me understand what the classNames do.
2. In-class Practice WODs
For the In-class Practice WODs I refrained as much as possible when it came to AI usage. I would only use it if I knew that it would be impossible for me to do alone. For example, in the React in-class practice WOD, I had no idea how to make a background image fit the entire screen, as well as had trouble making the navbar shrink when the tab itself shrunk. Otherwise, I did everything else myself. This led to a lot of DNI WOD’s, but I believe that the ability to work efficiently by myself under pressure was something I wanted to improve upon. Using AI as an easy get-out-of-jail-free card would ruin this experience.
3. In-class WODs
When it came to in-class WODs, I usually didn’t have to use AI at all to get the WOD done. Even though there were a few WODs that I DNI’ed on due to my refusal to use AI, I prefer attempting to work it through on my own. The difficulty of the in-class WODs varied slightly. Either it was easy enough to the point where I didn’t need to use AI or they were too difficult and I couldn’t complete it without simply pasting instructions into an AI model to do the WOD for me. However, towards the end of the semester, I did begin to use Claude to help me make minor adjustments to code, such as reminders of classNames and spacing.
4. Essays
I didn’t use AI for any aspect of the essays assigned in ICS314. While I believe in using AI for coding, that is because it is a very rigid subject with methods that absolutely work. Writing, however, is something that I believe to be deeply personal. Many essays ask for students to give their opinion on an essay topic and what they did. Even if you tell an AI what you did and thought, it will not be able to capture your feelings as accurately as you can. All writers have a writer’s voice, and AI takes this away from essays when you instruct it to write for you. This includes simple outlines, as you may have structured it differently had an AI not done it for you.
5. Final Project
I utilized AI for the mundane copy and paste types of code blocks. For example, when making an add-announcement-page and edit-announcement-page, I took lots of inspiration from the add-contact-page and edit-contact-page found within the Digits WOD. I gave Claude the announcement model and told it to edit the add and edit pages to match that of the announcement model. Afterwards, I would adjust it to my liking as well as to make it visually pleasing to admins. Additionally, I had difficulty implementing pagination as it was an outside concept that I used to reduce loading times when lots of objects existed. I used AI to help me understand the requirements and add pagination.
6. Learning a Concept / Tutorial
I frequently used AI like Claude when learning new concepts. For example, with React I had trouble understanding what each component did and how to use them. So I would give examples for Claude to explain to me, and I would then reaffirm my understanding. I would sometimes do this with WODs so I could understand them better. I would paste some code into the AI and ask about patterns and why certain blocks are code where in certain places
7. Answering a Question in Class or in Discord
I didn’t answer questions in class or in discord, so I never used AI for that. However, sometimes a good question came up in the discord that would never get properly answered. I would sometimes do my own research and use AI to help me find the answer. I wouldn’t answer, though, because I’m not very confident and I would not want to spread misinformation.
8. Asking or Answering a Smart Question
AI has answered many of my Smart questions during ICS 314. Whenever asking a question, I would try to give as much detail as possible about an error, including the before and after of the code block, the error message, what I’ve done to solve it on my own, and context regarding the structure of the code overall. By using smart questions instead of simply pasting an error code, I’ve gotten much higher quality responses with less back and forth between working solutions.
9. Coding Example (e.g. “give an example of using Underscore .pluck”)
Whenever I had to learn a new concept for the first time, it was often easiest to do so when I used AI to give me an example. I would take the example, and try to break it down on my own before going back and reconfirming my understanding with the AI. By doing this, I found it easier to then create my own examples and apply them to WODs.
10. Explaining Code
When using AI, I tend to focus more on explanations of examples I’ve found, rather than creating completely new ones. After understanding the base concept, oftentimes there were more complicated aspects I didn’t understand in the WOD or online. If I did not have enough time or had trouble, I would go to AI to explain what the code does and then I would compare that to what I already knew. Rather than thinking of examples and explanations as two separate ways of learning, I think of it as sequential, with explanations coming after understanding the basic examples generated.
11. Writing Code
While I didn’t use AI to write code for me often, I did use it to write some code for my final project. The final project used skills learned throughout the semester, so I knew most of what to do when it came to coding. So, I used AI to generate code that I already knew how to create, but would’ve been time consuming to type out. Of course, I would check and rewrite bits of code that I didn’t like or didn’t pass ESLint.
12. Documenting Code
I didn’t use AI to document my code. When documenting code, I find that I understand what I write best, so it was easier for me to just write concise notes, rather than use AI to write something and copy and paste. However, I noticed that Claude would document when writing its own code, I would sometimes change that to match the terminology discussed in my ICS 314 class.
13. Quality Assurance (e.g. fixing ESLint errors)
While I fixed most simple errors on my own, I found it helpful to use AI to help me make the final project aesthetically pleasing. For example, I would use AI to help me with the .css file of my final project when it came to coloring buttons and fonts.
14. Other Uses in ICS 314 Not Listed
Other than the methods listed prior, I did not use AI in any other way. I feel that the previous prompts covered the methods of AI usage that I participated in.
III. Impact on Learning and Understanding
Overall AI had a positive impact on my learning of concepts and coding in ICS 314. It was helpful to have a resource to help not only give examples, but also detailed explanations of very specific bits of code. It helped me understand coding in a way that I could control and pace. This is much more personalized than the videos and readings that are often too difficult for me to understand fully. Additionally, even though the videos provided alongside the WOD in ICS 314 were useful, oftentimes I felt like they moved way too fast, to the point where I would ask AI to help me understand what is being said there. This helped me keep up with the class and eventually allowed me to complete in-class WODs without the use of AI.
However, I noticed that halfway through the semester, I began to overrely on AI. I began to get lazy and simply prompt it to do the work for me. This decision affected me for a while, as it left a gap in my knowledge where I am weaker at certain concepts due to the ease of AI answering for me. I would be able to answer questions, but not understand why they worked or how they worked. This gap of knowledge reinforces my point of view that AI must be used carefully and cautiously in order to grow and learn.
IV. Practical Applications
I tend not to use AI in my day to day life. Outside of ICS 314, I do not find myself struggling as much on understanding concepts or designs. However, Asian language courses at UHM use something called “Moodles” to help reinforce learning. These moodles are made by the teachers, so there are natural errors and inconsistency due to lack of attention to detail. If I ever have to spend more than half an hour on a moodle, I will sometimes turn to an AI to make sure I’m doing it correctly and I’m not in the wrong. Sometimes I get the concept of the homework correct, but I forget a small mistake like past tense. So, AI can be helpful in checking my work in other classes.
V. Challenges and Opportunities
The main challenge of using AI in general in an educational setting is knowing how to use it to help you learn rather than help you complete assignments. While AI proves to be incredibly useful for getting assignments done in short times, the fault is that the user doesn’t end up learning from anything. In the long run, this sets users up for failure as they may find even the most simple code blocks impossible to recreate. However, the opportunity to have what is essentially a tutor available at all times makes it an invaluable tool when used correctly. Not only can it come up with examples of how to use code, but it can also analyze hard to read code and explain what each line does. This is especially helpful when working with partners that may use code you don’t know about. However, I believe that all AI answers should be taken with a grain of salt. Your vision and the AI’s vision may always differ.
VI. Comparative Analysis
Compared to traditional learning, AI has pros and cons to match. The major pro of AI learning is the convenience. AI is available almost whenever you are connected to the internet and will do its best to answer any question prompted. However, the validity of those answers may be questionable depending on the subjectivity of the question, which leads to incorrect learning rather than no learning. The major downside of traditional learning is pacing and the loss of individuality.
In traditional learning environments, everyone must learn at the same pace and those that cannot keep up fall behind. While this works for some people, it may be oppressive for others. On the other hand, AI is very structureless, requiring the user to want to learn rather than want to get a good grade. Even though this may work for individuals with discipline, the lack of structure may cause users to give up and feed the AI the questions to get answers and get it over with. Overall, both traditional and AI learning have their pros and cons that may work for different people.
VII. Future Considerations
I believe that until the AI bubble pops, if it does, that AI will become engrained into most parts of daily life. This has been proven so far as people go to AI for everyday questions rather than using google. Even though AI can give answers in a second, they are not always the best nor entirely accurate. AI works by taking what is known on the internet and regurgitating it, thus allowing for the possible spread of misinformation or lack of efficient solution. I believe that if AI is fine tuned to an academic setting, taking from specified research papers, it can be trained to be a reliable teaching tool in addition to traditional learning.
VIII. Conclusion
Overall, AI is growing larger and larger in regards to daily life. My experience in ICS 314 is no exception, with AI guiding me to understanding abstract concepts and assisting in code development. I used Claude to help me understand the code presented within at-home WODs and help me understand the solutions behind the in-practice WODs.
Even though it may seem at first to be an indisposable helper, it’s always important to remember that you need to know what’s happening. You can’t let your code become a black box that only AI knows how to decipher and write. Instead, AI should work to teach you why and how code works.