This question is one which appears often among communities I'm part of. Many fellow students share an uncertain stance on the use of AI in programming and that's very valid especially since many schools and institutions push the narrative: "AI is bad" or "You should not use AI or there will be consequences". Although it can be argued that this narrative is slowly shifting towards more acceptance of AI in coding, the uncertainty has stuck and many students are confused on how to use it in order to reach a perfect balance of learning and support.

My Personal Stance

Personally I think the educational narrative of "AI is bad" is flat out incorrect. I have used many different AI models to help me revise, refine ideas, and learn new concepts and it has helped me in each of those areas tremendously. A great example of this is an exam on certain topics and programming concepts in C which I had just over a week ago. My friend and I had just spent the week before trying to finish a 4,000 word essay and had very little time to revise so 2 hours prior to the exam we pulled up GLM 5.1 and prompted the model to provide us with 50 questions relating to lecture slides which we provided it. After 2 hours of going through the questions, researching topics we struggled with, and trying to cram as much as possible we felt very confident going into the exam. We both left the exam with a score of 86% each which felt great.

That wasn't my only recent experience using AI productively, I've also leveraged AI to help me structure the aforementioned 4,000 word essay and although I am yet to receive results for it I felt confident that it met the requirements of the mark scheme and guidelines. In fact I'd say it tremendously increased my productivity as I didn't have to concern myself with how to structure the essay but instead aim all my focus at writing code and the content of the essay itself.

My Recommended Best Practices:

In my opinion it's easy to cross a line using AI and have it write all your work and code but that's counterproductive due to a lack of challenge and actual learning. If you put no care or effort into it then you ultimately learn very little to nothing.

Firstly, you should rarely if ever have AI write an entire piece of work for you. Instead have it structure your essays or help plan out your coding tasks into easy to understand steps. This allows you to focus on the main body of your work and increase productivity substantially.

Secondly, always provide the model with your course handbooks and assignment briefs/guidelines. This way you don't waste time typing out a long prompt explaining the entire assignment and also provides the model with the exact and fully accurate information it needs to help you plan your work.

Thirdly, if you are revising for an exam always prompt the AI to teach you and quiz you until your quiz answers are adequate. By doing this you won't get walls of boring text and actually allow yourself to practice. Pairing this with the previous tip will allow you to revise extremely efficiently and you'll be surprised with the progress you make after an hour.

Conclusion

Institutions claiming "Using AI is wrong and dishonest" do not understand how beneficial AI can be if you use it appropriately and are ultimately falling behind. If you're a student yourself and struggle to find the right balance using AI models then the above advice should help you get there. Obviously feel free to experiment with different prompting styles and methods but the above tips have worked incredibly well for me.

Please feel free to let me know of any tips you may have in the comments and any feedback you may have for my tips.