I can confidently say that at some point in their college career, every engineer has been stuck on a tough problem and tries to copy and paste their homework directly into Google or ChatGPT. I am guilty of doing this myself, especially late at night when I just want to finish the assignment. Unfortunately, this strategy does not always work out the way I hoped it would… Other times, ChatGPT gets me unstuck, and I find myself with a deeper understanding of the content or knowing I had a misunderstanding before.
Of course, each Professor has their own opinion on AI, so it is important to be mindful of their policies. Yet, when AI is allowed, I find it to be a very useful tool that saves me time while actually increasing my understanding of the concepts.
I use AI in two major ways. As an Electrical and Computer Engineering (ECE) major, I do write a fair amount of code, but this is rarely the primary focus of the course. Often, in classes such as physics, circuits, or linear systems, we will write MATLAB, Python, or C++ code to perform theoretical simulations or complex calculations. I often will use AI to help debug my code. I wish I could say AI was able to write all my code for me (it would save me so much time), but usually, AI is unable to do the complex theoretical work behind the surface-level code and is mostly useful for debugging. I think that this is an excellent balance since debugging code usually means reading every single line, looking for errors. While I think this is a useful skill to develop, when you have 500 lines of MATLAB code to simulate the behavior of the antenna you are building, AI can be a great way to quickly verify your syntax before you spend too long checking your theoretical calculations.
The second way I use AI is to help understand complex topics. Again, I cannot say that AI is able to solve most of my problems, especially in upper-division engineering classes. However, it can be useful to solve small problems or help clarify a misunderstanding. I think if AI is given a large physics problem, for example, it will make several mistakes along the way that make it almost useless. Sometimes, I give AI a small section of my derivation and have it fact-check me, which it is quite good at.
I think AI can be a useful tool, but I also worry that I am too reliant on it. Whenever I can, I try to solve the problem myself first, since this is how I will have to solve them on the exam, and I will often consult my friends and classmates before consulting AI of any kind. However, I do think it is important to know how to use AI accurately and efficiently.
