There is nothing wrong with using AI to write code when you have the experience to guide it, evaluate its output, and recognize mistakes. Like operating a forklift, using a powerful tool safely requires knowledge and training. Giving someone access to that tool does not automatically qualify them to use it.
I use AI myself, and I accept AI-assisted pull requests when the contributor understands the code they are submitting, has tested it, and can explain their decisions. The contributor remains responsible for the quality and correctness of the work, regardless of how it was produced.
My concern is with people submitting code they do not understand and cannot properly assess. In experienced hands, AI is a powerful development tool. Without the knowledge to identify and correct its mistakes, it can introduce serious problems.
