How Can I Contribute to an Open Source Project Without Using AI-Generated Code?

A straightforward approach would be to avoid using any AI from the start. If the project's rules explicitly state no AI-generated code, you'll have to stick to that. As long as you can genuinely understand what you're writing and reimplement it from scratch, you should be good to go. Taking the core logic you learned and expressing it with your variables and structure can work well. If you’re just making minor tweaks, it might still be too close to the AI output for comfort.

The real test is your understanding. If you can explain the logic and rebuild the code without the AI input, you're likely in a good place. Maintainers want to trust the code they're receiving, so if you can demonstrate that your work is your own, you'll be respected. Just rewriting it blindly won't cut it, and you'll want to ensure it aligns with the project's coding style.

It's hard to draw a line on what constitutes 'using AI'. If you read AI-generated content online, does that mean you can't apply similar solutions from it? Given how integrated AI has become in development tools, it can feel impossible to completely avoid it. But if you're using something like an IDE that suggests completions, I think that’s fine, as long as you're ultimately writing your own code.

Honestly, if you can write it from scratch without referencing the AI model and still maintain the logic, then you're likely creating a true "human" contribution. Seek to rewrite and deeply understand the processes instead of just copying what's provided. This way, you're contributing ethically and respecting the original maintainers' wishes.