In recent years, AI (artificial intelligence) has become more prevalent due to ease of access in programs such as ChatGPT, Bard, Copilot, etc. AI has been used to generate art, essays, code, and many other things depending on the user’s imagination.
The use of AI has become more controversial due to the lack of regulations on those who use it. In the art scene, many artists are frustrated since AI programs get trained on existing work in order to generate new images. This has led to conflict because those who use the AI programs to generate images are able to do so without permission from the artists whose work is being trained on.
AI should strictly be used as a tool to help with templates to work off of, otherwise you are essentially plagiarizing existing work. If AI programs are too heavily relied on, it will lead to a long-term decrease in ability to problem solve issues as well as a lowered understanding of the generated product. For the students that used ChatGPT to fully write their code, there is a higher likelihood that they would not understand the generated components, nor how to adjust it to the specific project.
In the class ICS 314 (Software Engineering) that I took this semester, I have not used an AI program for a single assignment (including WODs, essays, or any of the code that I have written this semester). Instead, I have opted to work off of templates that have been either provided by the instructor or my own past assignments from ICS 314. This forced me to understand the relations between the components and elements within the templates and allowed me to adapt it in my own way.
Experience WODs are typically multipart homework assignments that require the student to build off the work from each of the previous parts. This was one of the elements of the course that I learned the most from. I chose not to use AI for this element because I figured it would be counterintuitive to let the AI do the work, considering that the value of the Experience WODs was in personally coding it from scratch.
On Mondays, our class would go through In-class Practice WODs. These would be set up to be similar to the graded In-class WODs on Wednesdays. Similar to the Experience WODs, I found these to be valuable to learn the content, which meant that using AI would just hinder my experience.
Throughout my various essays that I have written for this course, I’ve expressed my disdain for the grading system for the In-class WODs. I felt like they would have been more valuable to my learning if I didn’t need to stress about an all-or-nothing grading approach. I was the most tempted to use AI for this portion of the course, but chose not to because I did not have enough experience utilizing AI. I figured in a timed setting, using unfamiliar tools would end up working against me. I felt incredibly frustrated because all of the WODs that I did fail were very close to passing and only needed a few tweaks to get it to work. Use of AI in this situation would have actually been optimal since it would have been used as a reference tool rather than to fully generate the code.
While I’m not a big fan of writing, I don’t actually mind personally writing essays since my personality and ideas show very clearly in my writing. I chose not to use AI for essays because I felt that it would cause me to spend more time in the end changing the writing to actually match my own tones.
The Final Project is a group project and requires the cooperation of all team members. I feel like AI can be a great tool to utilize in this aspect as long as the use is small in scale. If the use of AI is used on a large scale, it will cause too many unknown issues that will require more time to debug than if we were to write it from scratch. I personally won’t use AI for the Final Project due to my lack of experience with it.
The point of learning a concept is to go through the motions yourself. I learn the best from a hands-on approach, which means that the use of AI here would be detrimental.
I only choose to answer questions in class/Discord if I feel confident that I can personally contribute to a solution. There is no point in using AI myself to answer someone else’s question since they could just use AI themselves.
Similar to my response above, I only choose to respond to questions that I can personally contribute to. Using AI would be disingenuous and could lead to more confusion since I wouldn’t have come up with the prospective solution myself.
I typically just Google for similar examples of code that I want to utilize. I would not mind using AI in this situation, but I cannot confirm the accuracy or effectiveness of the AI’s solution like I would be able to from a Stack Exchange forum. In a forum, I could use the responses to the “correct use” of the code to understand if it will be helpful in my specific situation or not.
I normally will either Google the explanations for code or check the documentation. The documentation is the most reliable since it tends to relate directly to the content that we are learning at the time. I choose not to AI here because I find it more accurate to have a human explain something than AI itself.
I think using AI to generate a template would be okay, but anything further would be too much. I like using the provided templates since it is always more relatable to the assignments. I choose to write my own code for a better learning experience. I could see it being more useful in a work environment under the right circumstances and proper monitoring.
Code documentation requires a very specific understanding of the written code. I would not trust an AI to understand my thought process, especially since I tend to write unconventional code. I think it is best to stick to manually documenting your own code rather than rely on AI.
There are already systems like TestCafe that automatically check the quality and functionality of your code. I think with the proper set up, AI can be a great tool to use in quality assurance. I have not used any specific AI tools for quality assurance in this course, but I would not be against doing so in the future.
Considering that I didn’t use AI at all in this course, I can say for certain that I did not have any other uses for AI in ICS 314 that has not been listed above.
I struggled in the first half of the semester because of my lack of initial knowledge in things such as JavaScript, JSFiddle, IntelliJ, React, Underscore, Meteor, and anything else that we utilized in this class. Looking back, I feel that I was only able to familiarize myself with these tools specifically because of my aversion to AI. I believe that AI would have helped me skip a few steps to solve my problems quicker at the expense of personally learning how to do it myself.
Course elements such as Experience WODs were only so impactful in my learning because I was able to utilize my own problem-solving skills. This ended up helping me in later In-class WODs because I got faster at troubleshooting issues. Using AI to problem solve anything would have meant that I would need to rely on it more in harder content which would make me a less valuable asset in a workforce. I believe that problem-solving is one of the best abilities anyone can develop for a career.
As a friend to many artists, I am strongly against the use of AI in the art space. In general, I think it is okay to utilize AI for templates to assist your work or to help troubleshoot. An over-reliance on AI will just lead to less personalization and be akin to plagiarism.
Humans do have limitations on their accuracy and computation, so use of AI in high end simulations can be useful but should still be heavily monitored. There should always be a human touch to a product to ensure the quality. At the end of the day, AI still needs to be created by a human, which means that they can also be subject to errors. There is no such thing as a perfect person or AI, so the cooperation between the two can lead to better outcomes.
AI is already implemented in many aspects of our society such as the use in agriculture to predict yields as well as monitor crops and pests. Since AI is being used as a tool in agriculture, it enhances the outcome rather than outright replacing all the work for it.
I chose to turn down any opportunity of using AI in this course and did not encounter challenges in utilizing AI because of it. My grade did suffer due to the lack of use of AI for In-class WODs however, but I prefer to have a lower grade in exchange for a better learning experience. Of course I would have preferred the better grade I could have gotten if the In-class WODs were graded on a scale rather than 0 or 100 points.
In the future with the proper set up, I believe courses could implement AI to help generate the templates or presentations for the students to utilize and learn from. I think there should still be some level of regulations for AI use to encourage students to learn on their own. I believe it would lead to a better workforce if more students did not rely too heavily on the use of AI.
The integration of AI in teaching methods will vary based on the professor teaching the course. If the professor is able to use AI as a tool to improve their course structure, it can enhance the learning experience. Aspects such as engagement, knowledge retention, and practical skill development rely more on the creativity and skill of the professor and cannot be replaced by an AI. Since the AI relies on prompts to generate something, you cannot create an engaging class out of nothing.
AI will likely become a helpful resource in the future of software engineering education if it is utilized to create things such as learning applications. However, professors will need to be clever to detect the use of AI by students to fully complete assignments since that would be akin to plagiarism. With the rapid development of easy to access AI programs, it will be harder for professors to completely prevent students from solving everything with AI. Courses will likely need to be adjusted specifically with that in mind (and already have been in recent years).
Regardless if you want AI to be integrated into society or not, it is inevitable that it will be further developed and available to everyone. Social media applications are already integrating AI chatbots and ChatGPT is the go-to application to generate completed assignments for many students. It is most important to develop guidelines for AI use to prevent the over-reliance on it in the future. Rules and regulations can only go so far, the rest needs to be monitored by each individual person. If we want to avoid a future where AI becomes more valuable than actual humans, we need to make sure that we continue to develop ourselves accordingly.