In the context of a student writing their first programs, the most important part of developing code is at the beginning of the development process: thinking of the solution. This may also be the hardest part of the process for inexperienced coders. This is also where students are most tempted to generate a solution using AI.

I learned the most by thinking about how to solve a simple problem as a human, then breaking it down into parts that I could implement in code. Once I had the idea, I could continue to implement and debug.

This brings me to the actual hardest part of coding: debugging. Although I just said that thinking of a solution is the hardest part of developing a program, debugging is the real challenge. It is crucial to develop your problem-solving skills in this aspect. Debugging is easily the most time-consuming part of developing a program because even if you debug your program and can find the exact line of code causing an issue, adjusting your program to fix it can be very difficult.

Personally, I find this is the best time to use genAI when coding. First, use debugging tools such as print statements to identify where the bugs originate. This is a very time-consuming process, however, this process helps exercise your problem-solving skills. Seeing processes in your code that function correctly and those that don't help you narrow down what code may be causing issues.

Rather than sending AI a prompt to complete (sometimes the code AI generates from scratch can be rather buggy as well), I paste a small snippet of code I know is causing issues and the AI application I'm using sends me back a list of potential solutions. This helps ensure that the specific problems I want to focus on are resolved. AI responses with a smaller number of solutions can also be easier to understand, and minimizing the number of changes implemented at a time prevents code from breaking in large, difficult to understand ways.

Searching for issues in code is a very important skill, but generating solutions using AI provides better learning for inexperienced coders than many other methods. Generated solutions often describe the logic behind why a certain line of code doesn't work, which students can remember and use in the future.

This method is far better for me personally also because by doing this, I have already undergone the process of developing the code and looking for the problem. For an inexperienced coder, generated solutions can also provide early exposure to best practices in coding style (such as SOLID (use object oriented programming principles (when applicable) that is modular and maintainable) and DRY (use reusable functions to reduce the amount of repeated logic in code), which are 2 software engineering methodologies describing the best practices for software coding style) before they learn about it later in classes.

When debugging with AI, you may have problems with your code that may require large amounts of rewriting. In this situation it can be tempting to copy-and-paste a generated solution and hope for the best. However, if you don't fully understand a solution and why it works before implementing it, you can miss some of the benefits behind coding with AI, and miss some of my points in the last few paragraphs. Instead, look at code that has been generated, and write it yourself. Try not to just copy the code verbatim line by line, but understand the logic and use your knowledge to implement it yourself.

While you implement a solution, if your solution isn't working, you can also look back at your generated solution as reference. This can be very helpful in understanding and learning from your mistakes, and also speeds up the debugging process.

"My staff teach these students how to use AI effectively to enhance their learning, rather than relying on AI to do the learning for them." - Laura Hollingsworth, Director of the Quantitative Skills Center UWB

I think that Je's point on AI's impact on human creativity is one that isn't commonly discussed - being about AI's impact on society as a whole, and how that will impact people, in a more subtle manner beyond AI's use to generate content.

This quote stuck out to me because it emphasizes the forward thinking of transitioning students from the classroom, which is ultimately the point of college which is to teach you the necessary skills and knowledge needed for the workplace with AI being the latest addition to those skills.

- Erik Echols, Writing and Communication Center Director

I chose this quote because losing marginalized voices isn't a common talking point for the argument against using AI for writing, and it was eye-opening when Erik pointed this out.

I chose this quote because I believe it illustrates how AI is a tool that can do both bad and good, but will eventually forward society and technological advancements.

✅ Makes my code cleaner
✅ Saves time
This is the kind of stuff AI is great for—it helps me refine my code and learn new techniques along the way. I used similar AI-assisted optimizations when implementing AES-256 encryption
and OAuth 2.0 authentication in my security projects.

More specifically, AI is an excellent tool for keyword or idea generation, and first-draft generation. Users should understand that AI has its strengths in summarizing or taking points from existing information, which is especially helpful in the beginning parts of creating some creative work. This provides something that removes some work from the user, but still requires the user to edit, and have their own creative or logical process. A first draft can be anything generated based on a user's general ideas or points for a creative piece. A generated first draft will then have all of these points articulated in some form, which will allow the user to edit it, adjusting vocabulary, flow, etc., which may be difficult for AI to generate in a unique manner.

I think this is an important concern to call attention to because there has been so much fear mongering around AI replacing the workforce. It’s a valid concern especially in education. I agree with what Reimer says. As a student, I don’t want to come to class to talk to an AI on a screen, because face-to-face communication is really important. There is a grey area for us to define where AI isn’t the primary method of teaching but where teachers can use AI to reinforce student learning.

Education Systems can have issues with valuing answers over students effectively learning new information. It is of some concern that AI could exasperate valuing these answers over students developing understanding.

Dr. Hollingsworth explained, “Students are still developing their knowledge and may not be equipped to identify what’s correct or incorrect from AI-generated answers. Within our tutoring center, students often use AI for problem-solving, coding, or writing assignments. While there are ways to use AI effectively as a novice, you need to be more careful. Learning requires engaging your brain without relying entirely on something else to think for you. Using AI as a tool to check your work or debug is fine, but if students let AI do the work for them, they’re missing valuable learning experiences.”

This concern is important to me because as a student I’ve sometimes felt that I’ve had to choose my grade over my understanding of a subject, and that while I was able to pass a class that doesn’t mean I remember it well. It’s also a concern of mine that students struggling with the pressures of school will use AI to get by and lose out on even more learning. Essays especially are one of the assignments that helps me learn new things and sticks in my head, and something AI could replace when answers are valued over learning.

I chose to highlight this practice because it is one of the best ways for students to maintain control of their work while still using AI. After going through multiple iterations, the generated content will better represent their voice and still promote critical thinking skills even when relying on AI.

It is interesting to think about how in the future AI will keep getting better and better, but at the same time, these advancements are most likely going to be blocked by massive paywalls and subscriptions that gives a hefty advantage to those who don't have paid access to higher level LLM's than those who does. It is very concerning to think about, but at the same time all we can do is wait and see what happens in the future.

However, its answers have to be sorted through and read carefully. I can't take it at face value and often have to reword questions to find the information I actually need.

It is concerning that students rely on AI for reflection prompts because these responses are meant to be personal and authentic. If AI is reflecting, we lose the chance to truly understand our thinking and growth.

This idea functions primarily because the students are past the fundamentals of computer science where AI has the most risk of causing detriments to learning.

Rowland explained to us how working as students, faculty, and staff on the AI policy for the course was a helpful process:

“...the sort of democratic process that we engaged in in that 476 class is a useful way for students to come to conclusions together about how to use AI in a class, and what sort of the boundaries are and what the expectations are, like agreeing upon those things as a learning community.” 

I choose to highlight this practice because especially in the later years of college it’s good to prepare students for how tools will be used in the workplace, which is often through discussion and collaboration.

I'm highlighting these practices because it shows a good mix of thinking about AI in education. Laura sees the good side of AI but also makes sure to stress that integrity is important. This seems like a solid way to use AI to help students without messing with learning standards.

 UW Bothell Professor May Lin in her interview with Travis Huynh talks about how she has changed assignments to be more collaborative and personal to promote engagement and thereby [verb] less use of AI

She encourages other educators to take this approach: “So make it more personal, that they really do want to invest their time in working on the assignment instead of just doing it, so they can, you know, go through the hurdles. “

I choose to call attention to this practice because I view AI as a rising technological force that should not be resisted but rather adapted to. I think that the high use of AI for assignments should be considered as feedback for the professors that maybe their assignments don’t feel personally rewarding or worthy to put effort in for the vast majority of students. I think revising assignments not only promotes learning but also helps fight the loss of critical thinking due to high use of AI. 

A key piece of evidence supporting this comes from UW Bothell Instructional Designer Todd Conaway’s statement: “Part of my background is like a poet. So I was a high school English teacher. So I really liked the words. And I like words that I make into something like that's important to me. I know that's not true of everybody but to me my ability to move words around and make them make something is really important to me and so, I just would never give that up to some freaking computer thing.” This shows a concern that AI, while powerful, could lead to absence of personal and human aspects of creative expression.

This question is important to me because it highlights an important tension between leveraging AI's capabilities and preserving the unique, human-driven and human-thought aspects of education. If AI is to enhance learning, it must be thoughtfully included into education in a way that supports, rather than replaces, students' and teachers' creative processes. This balance is something to be found and established as we shape AI's role in classrooms.

“If we give up our responsibility to create sentences and patterns of words in our own heads, each one of us in our own heads. If we give that up to some other tool to then represent our thoughts to the rest of the people around us. We'll actually become less able to use words well.”

I believe that he brings up a strong case for concern. With AI becoming more and more prevalent and people using it more, if it’s not used in the right ways it can have negative effects on people. As expressed from interviewing with Tood, his main concerns was about peoples creativity and voice. The human mind is able to come up with many incredible ideas and things and it would be a shame to always turn to AI for inspiration or other ideas. Todd claims, and I agree, students and other individuals are capable of doing amazing work. We are capable of so much more with and without AI but AI shouldn’t become a crutch or go-to tool people are constantly turning to. Todd’s concern speaks to me because it raises the awareness of complacency and laziness AI can have on people if not used properly. It highlights the way AI should be used and how to approach using AI to make a positive impact. 

Nelson explained to us how he is working on incorporating AI in his classes so students get practice with it. 

“So I've developed a general pedagogy for using and teaching AI in the classes that I do. In class itself I will teach using the tools, demonstrate it and then have them practice and try it out. Then they're kind of like required just to try it out, just give this a try.”

I think one of the most important practices we heard about from educators with AI and how it’s advancing so rapidly, is learning and building your skills at how to use AI. These are fundamental skills that students should know how to do or be able to learn how to do. Educators should give students opportunities to try AI in their courses. You don’t necessarily need to use AI all the time, but the ability to understand how to use AI and having actual practices with it is important. Like anything else, practicing makes you better and the same can be applied to AI. Practicing and exploring how to use AI and find out what its capabilities are is an important practice.

AI has significantly changed my coding journey. It has changed me to the point where I might not need to write code from scratch again for the rest of my life. I use AI to write most of my code and the majority of my time is spent on refactoring and debugging the code. 

How I use AI in my coding

Fixing Bugs, Not Writing All Code

I used to write every line of code by hand. Now, it has been years since I wrote code from scratch. Now my focus has shifted from writing code to debugging it. AI will write most of the code and what I basically focus on is just debugging and fixing that code so that it’s actually doing the thing that I intend it to do.

Building Full Apps with Little Code

I have made full web apps with AI. AI writes most of the basic code, then I change and add what I need. This way, I can make a working app without having to type out every single line myself. My imagination and prompting skills are much more important than my coding skills to build different applications and websites. With this approach, I've successfully built three full-stack web applications. All I needed to do was write very good prompts and focus on fixing the code to create the functionalities that I intended to build. Additionally, modern tools like Bolt or Lovable allow entire web and mobile app creation using just prompting and a single deployment button, significantly simplifying the development process.

Planning the Design

Initially, I let AI handle the design entirely, which led to messy outcomes. Now, I plan the big picture first, then use AI for smaller components. AI can sometimes stray from the original plan or omit critical pieces, making debugging difficult. Therefore, I never let AI manage the initial design entirely. Effective initial planning helps minimize costly debugging.

Vibe coding

I recently came across the phrase “vibe coding,” perfectly describing my everyday use of AI. I ask AI for code, test it, and if it doesn’t work, I feed the errors back into AI. I repeat this cycle until achieving the desired result. Although convenient, sometimes simpler errors are quicker to fix manually.

When I Don’t AI

• I never let AI completely handle design decisions. I always participate actively.

• When AI proposes unnecessarily complicated solutions for simple fixes, I prefer coding myself.

• If a fix is straightforward, manually addressing it is faster than waiting for AI.

A workflow of My AI Use

1. Identify the desired code or the issue in existing code.

2. Send AI a brief message describing what code I need or the problem to fix.

3. Review AI’s response to ensure clarity and correctness.

4. Test the AI-generated code to verify it meets expectations.

5. Repeat the process until satisfied.

Conclusion

I believe coding should be simple, clear, and eventually done in human language. AI greatly aids in starting projects, writing code, debugging, and overall efficiency, positively transforming my coding experience. However, it's crucial to actively participate in design decisions and avoid completely AI-generated code, as it can lead to complex debugging issues.

This approach is important because it helps students feel more connected to their work. Instead of just doing assignments to get them done, they see the value in what they’re learning. It also helps them apply what they learn to real life and other classes.

I chose this practice because it focuses on what students need and enjoy. It makes learning more meaningful and helps students stay motivated, which is especially useful in a world where AI can easily complete generic assignments.

This matters because education should be fair for everyone. If only some students get access to better AI tools, it could make learning harder for others and widen the gap between them. As AI becomes a bigger part of school, we need to make sure all students have the same opportunities.

AI in Action: Generating Boilerplate Code

Boilerplate code refers to sections of code that are repeated in multiple places with little or no variation. It is often necessary to include for a program to function properly but doesn't contribute to the core logic. Boilerplate code is common in languages that require a lot of setup, such as Java, where you often have to write standard class definitions, import statements, and method signatures.

Imagine I am working on a Flutter app that requires user authentication. Instead of writing out the entire Firebase authentication flow from scratch, I use AI to:

• Generate a structured Firebase authentication setup in Dart.

• Create API request functions with appropriate error handling.

• Scaffold UI components such as login and signup forms.

• Ensure consistency in repetitive sections of the codebase.

Since I already know how authentication works, I can confidently review AI’s output, modify it where needed, and integrate it efficiently into my project.

When Not to Use AI

While AI is great for speeding up my workflow, I avoid using it when I do not fully understand the code it generates. Over-reliance on AI in these situations can lead to gaps in knowledge, difficulties in debugging, and issues when AI-generated solutions do not align with project requirements.

1. When Learning New Concepts

If I am learning a new algorithm, framework, or language feature, I avoid letting AI generate the entire implementation. Writing code manually helps me grasp the logic, which is crucial when I need to debug, optimize, or explain the implementation.

2. When AI Generates Code I Don’t Understand

If I find myself copying and pasting AI-generated code without fully comprehending how it works, I take a step back. Instead, I break the problem down and write the code myself or research relevant documentation.

Conclusion: AI as an Accelerator, Not a Replacement

AI is an excellent tool for generating boilerplate code efficiently, but it should not replace fundamental coding knowledge. By using AI to handle repetitive tasks while ensuring I fully understand my implementations, I can work more efficiently without compromising my ability to debug and adapt solutions as needed.

I chose this practice because it helps students develop essential skills for the future. AI is becoming a big part of education and work, and knowing how to use it properly is important. If students learn to question and understand AI-generated content, they will be better at thinking for themselves. This also prevents them from depending too much on AI and missing out on real learning.

I chose this practice because it helps students develop essential skills for the future. AI is becoming a big part of education and work, and knowing how to use it properly is important. If students learn to question and understand AI-generated content, they will be better at thinking for themselves. This also prevents them from depending too much on AI and missing out on real learning.

This is important to me because many assignments that haven’t been updated since the widespread adoption of AI tools like ChatGPT have become obsolete. If educational materials are not updated with AI in mind, students will continue to use it in ways that harm their learning.

Todd Conaway, one interviewee comments on an application of AI tools that uses summary, which provides information from large amounts of information. “You search something on Google and that little summary you get right off the bat, that's AI generated and, I think there are a multitude of ways that it will be helpful to us. And it'll be helpful exactly like the automobile.” - Todd Conaway. 

Summarization is one application of AI tools that AI is well-suited to. Todd explains AI summarization for search results is one of many helpful applications. Although many people have fears related to AI, how it can used, and how it will change society Todd emphasizes that AI is a tool like any other, and is very helpful when used in as such. However, in order to use AI most effectively, it is important to understand what applications AI does well at and which ones it doesn’t.

As another example of an application AI is well-suited for, Je Salvador recollects their use of AI to generate research keywords, stating “[AI was] giving me these key terms that I would look up and realize that it had these entire cultural references and access to those that I wouldn't have ever known about.” - Je Salvador. 

To support AI literacy, it is especially important to understand what applications AI have strengths in. This allows users to use it to its best advantage, and also provides users transparency on its weaknesses. Awareness of AI’s best applications prevents users from being overly-reliant on AI tools.

This is important because some students use AI in ways that limit their learning, like generating personal reflections instead of sharing their thoughts. Schools should help students use AI as a tool for learning, not as a shortcut that replaces thinking.

I chose to highlight this practice because it presents a promising way to integrate AI in education without diminishing students’ intellectual growth. As AI becomes more prevalent, educators need to design assignments that ensure students actively develop skills rather than passively accept AI-generated content. Lin’s approach serves as a valuable model for creating learning environments that balance AI use while preserving essential human cognitive development.

This approach makes a lot of sense. AI is a powerful tool, but without reflection, it can become just a shortcut. By thinking about when AI was helpful or misleading, students develop stronger analytical skills and digital literacy—both essential beyond the classroom. Reflection helps them recognize AI’s limitations and avoid over-reliance.

It also acknowledges how students actually use AI rather than pretending they don’t. Encouraging structured reflection ensures AI is used thoughtfully, keeping learning meaningful and helping students stay in control of their own intellectual growth. This balance between technology and critical thinking is key to responsible AI use in education.