I recently finished my first year of a bachelor's degree in applied informatics. My previous education was in chemistry, so I started with far less programming experience than many of my classmates. During the year, several instructors encouraged us to generate code from prompts and then adjust it until it worked. When I asked one instructor for help understanding a problem, I was simply told to describe it to an AI tool and let it solve the issue.
I understand that AI is becoming important in the software industry, and I know learning to use it effectively is a useful skill. However, I worry that I am learning very little when the tool does most of the thinking. I want to understand programming fundamentals, debugging, and problem-solving well enough to work independently rather than blindly accepting generated code.
In my spare time, I have been studying Java with Head First Java and building small programs. I occasionally use AI to explain a difficult concept or help with a problem I cannot understand, but I try to write and reason through the code myself. Is this level of AI use normal in programming education today? Should I simply follow my instructors' approach, or is it sensible to keep building the fundamentals independently and treat AI as an occasional learning aid?
5 Answers
AI is becoming common in professional programming, but it cannot replace the fundamentals. You need enough knowledge to judge whether its suggestions are correct, choose an appropriate design, debug the result, and recognize when it has invented something. If you cannot do those things without the tool, you will eventually get stuck. Keep studying the basics and use AI as assistance rather than as the person doing the assignment.
You are not wrong for studying outside class. Coming from a non-IT background means you may need more time with syntax, data structures, debugging, and software design than your course assumes. Keep building small applications, solve problems without assistance first, and then use AI or documentation to investigate what went wrong. That combination will prepare you much better than relying on generated code alone.
A good way to use AI while learning is to make it act like a tutor. Ask for a line-by-line explanation, request small hints instead of complete code, ask why a particular approach works, or have it suggest tests for your own solution. Type the code yourself and make sure you can explain every part. A textbook and hands-on practice should remain the foundation.
There are reasonable assignments where AI is deliberately used to demonstrate its weaknesses, followed by an oral explanation or code review. That is very different from an instructor responding to every question with 'ask the AI.' The latter is a warning sign because AI can produce confident but incorrect answers. Learning to use it is worthwhile, but learning to think and program independently is still essential.
Your instructors' approach sounds more like replacing teaching than teaching students how to use a tool. Asking AI to explain an error, compare possible solutions, or review code you wrote can be useful. Asking it to generate the whole solution and then pasting it in mostly teaches you how to prompt. Continue with the book and small projects, and try writing the solution yourself before asking for hints or explanations.
That is what I was worried about. I will keep working through the fundamentals and use AI mainly to clarify concepts or point out mistakes rather than generate everything for me.