Is AI a blessing or a curse when it comes to learning?

27 Apr 2024

I. Introduction

the realm of modern education, where the lines between human intellect and artificial intelligence blur, the classroom experience takes on a futuristic hue. Funny enough, the first line of this essay was written with AI, but it’s correct. ICS 314 is the most AI positive course I have ever taken, and for better or for worse, it has made an impact on my outlook on the future and how I currently do my daily work. In terms of education and honesty, I like to use AI to check my work. The real skill in using AI is learning whether or not the answer makes sense for the situation. As the user, you always have the power to decide whether or not to believe what your AI of choice spits out. In my personal use, I gravitate towards chat gpt (3.5). I have tried others, mostly copilot, but for the purpose of this course, I feel that chat gpt gives the most throughout responses. I tend to like copilot if I want to know where the information came from because it cites sources, and it’s great to be able to validate where the AI got it’s information. With chat gpt, it doesn’t have this feature, but I have found that it gives the most detailed responses and is less likely to give the same result twice since it tries to adapt more. I like to ask chat gpt for clarification on it’s original answer for better understanding and I find that this helps. But let’s get into specifics.

II. Personal Experience with AI:

I have used AI in class this semester in the following areas:

  1. Experience WODs e.g. E18: I tend to use chat gpt to help debug my code with experiences. With the experiences, unless I am in a position where I have tried to solve a bug for a while without success, I avoid AI for the experience WODs. I felt it was a good use of my time to practice the “old fashioned” way for a situation, like an interview, where you can’t use AI. And it helps build skills for the real WODs.
  2. In-class practice WODs: Again, I didn’t use AI for the practice WODs because it was easier to work in a group and not use AI and it was good practice.
  3. In-class WODs: I very often used AI for the in-class WODs. Simply because they were timed. This was mainly for the first few WODs where I would just copy and paste the directions from the problem into chat gpt. If we were expected to understand how to do the problem from the instructions (which were very detailed), I didn’t see why I couldn’t handle it. Most of the time I would give some context (for example, I would say that I was using react to design a website and then I would give my code that I had been using as an example for it to follow) and then asked for it to complete the task. When it came to the later WODs with HTML and Meteor, it got harder to use AI with so many moving parts. At that point, my AI use fizzled away more. I still used it for some debugging, but with AI not knowing all the context of the code and trying to get it all the information, it was a waste of time.
  4. Essays: I didn’t use AI for most of my essays because the writing feels flat to me. I used AI for the last essay because frankly, I couldn’t figure out how to put my thoughts together in a nice way. So the way I used it was that I wrote out a messy essay and just told chat gpt to make it better. I have also used chat gpt for grammar corrections.
  5. Final project: For the final project, I have used chat gpt a lot for debugging. I had a lot of issues with deployment and I was also using chat gpt to help understand the documentation of the library I was using for my calendar page. I had to go get some help for the deployment (even though gpt helped for some of it) but for getting started with implementing the calendar, it was helpful to know what I needed to do to make it work without reading pages and pages of documentation.
  6. Learning a concept / tutorial: I have used chat gpt for this, but usually I would just ask “how would you explain this concept” or “can you put this in a way that is easier to understand?” and with a combination of this and using google to see if reading other material makes sense now, I would say it works.
  7. Answering a question in class or in Discord: I have used chat gpt for this, but only surface level to see what it would say.
  8. Asking or answering a smart-question: I didn’t use AI for this as I didn’t answer any questions and I felt that I would have to provide too many details to chat gpt about the smart question that I might as well ask the question myself.
  9. Coding example e.g. “give an example of using Underscore .pluck”: Yes, I have used chat gpt for this too with the exact example from the rubric for this essay. It gave a good perspective in the context functions could be used.
  10. Explaining code: I am sure I used chat gpt for this, but I can’t remember any specific time for this or details about how I did this.
  11. Writing code: I typically used chat gpt to write code and take snippets from examples. I tend to give far too many instructions but it gives good results. See more about this in the in-class WOD and final project sections.
  12. Documenting code: I haven’t used AI for this as I really haven’t documented my code, unfortunately.
  13. Quality assurance: This is probably my most frequent use of chat gpt because it gives the best results.
  14. Other uses in ICS not listed above: None that I know of!

III. Impact on Learning and Understanding:

When Google was first released, many “haters” thought that people weren’t going to be able to research anything and that they were going to loose important skills. While this is somewhat true, it made information easier to find and more accessible. I would say the same for AI. It is weird because I first heard about using AI 1 year and a half ago when it first came out, and to think it’s almost been 2 years and AI is having THIS big of an impact on the world. It is crazy. And it’s only going to get crazier.

In terms of how I am learning, I feel as though I have noticed a pattern of my study habits and use of AI from high school to college. I think I can somewhat attribute this to social media addiction, but my attention span has gotten much shorter and I get bored faster. With AI, it certainly helps when AI can help summarize or find the information I need for my specific problem to save me time. It is still something that I think I need to work on as a whole and not just specifically with this class (improving my study habit and attention span). I find that AI had made things a lot easier for me to understand and in way, challenges my thinking as I try to validate it’s claims. I like to treat AI as like, a tutor almost and a conversation. I am quite shy when it comes to getting help for course work and don’t want to spend the money on a private tutor, so chat gpt has helped me bridge the gap in a sense. But I do think my learning could be improved with out it because it does just give you the answers (if you decided to believe it).

An interesting observation I have made with AI is that it gives me more confidence to try something new. When developing the app for the final project, I really wanted to stick to things I already knew about but once I remembered how helpful AI was in case I ran into issues, I decided to give it a try. I think it’s a positive thing that comes with it.

In specific relation to software engineering, I think that AI has both helped me understand concepts better and has given me a way to be lazier when coding when you can just ask it for the answer to a specific coding problem. I think being able to ask AI for examples with how to do concepts and having AI there for debugging situations saves so much valuable time, but asking chat gpt how to do the problem does take away from the learning. If I couldn’t do the problem without chat gpt, it will be impacting my ability to answer something like that during an interview.

AI is a double edged sword when it comes to education, and it will be interesting to see how it plays into the real world of software engineering when we get to that point.

IV. Practical Applications:

Funny enough, nothing could come to mind for real world examples of this because I have never actually heard of a business openly saying they use AI in a specific way (other than to sort through job applications). So of course, on theme with this essay, I asked AI for some real world applications. It seems that the really big companies like Google, Microsoft, Facebook, Amazon and Tesla are all using AI for their software development! Not only do they use AI to develop software, but they also are using AI in their products. For example, Tesla is using AI for data processing, sensor fusion, object detection and more in their self driving software stack. It seems like it is doing well for them since these companies are very obviously successful.

V. Challenges and Opportunities:

AI is not perfect, and as I have mentioned before, AI has some issues of overconfidence. AI, if convinced enough, would tell you confidently that 2+2=5 until you correct it. And even then it might not give be best answer. This is a major challenge that AI needs to be able to overcome. I wasn’t able to get the AI to work with intelliJ, but I think that would be super helpful too so that AI could have access to all your files for the app and be able to give suggestions about it. I hope there is a course that goes into the relation between AI and software engineering or just ways to make use of AI in a smart way. I think these would be great ways to further software engineering and AI education.

VI. Comparative Analysis:

I feel like I have mentioned this before, but although AI makes the work easier a lot of the time, the knowledge retention is usually worse than if someone learned something by not using AI. Then again, I typically find myself using AI to do something that I have little experience with. I don’t find myself asking AI the same questions on how to do something, meaning that I probably have learned something from doing this. The technology is so new though and there are other outside factors in this age that possibly also contribute to learning of new information, so only time will tell what the real impacts are with using AI in the classroom and for learning.

VII. Future Considerations:

I think AI will continue to grow to help the ways of humans and hopefully also challenge us to become better as well.

VIII. Conclusion:

I think that using AI in this course was very interesting and being open about the use of AI was very refreshing. From what we have seen in articles and online, a lot of students are being punished for their use of AI in the classroom, which in certain cases makes sense and in others might be too strict. It is just another tool we can use, but like our tools, it can also be abused. Careful use of AI in ways that won’t inhibit learning are very important in teaching. By far this has been one of the most interesting courses I have taken because of the encouragement of using AI and I appreciate us being able to be educated to use this technology responsibly and to experiment with what works and what doesn’t.