How much should AI be involved in an into to programming course?
I am considering pitching a class as an adjunct professor at my local college. I want to teach introduction to programming (JavaScript) while teaching students how to automate Google Sheets.
I want the course to help my students build software that could change their life or view of what's possible in the world, while also giving them some real skills.
How much should AI be involved? I see three levels:
**No AI at all.** This is how I learned to do it, and I know writing the code out by hand yourself is really important for teaching you how to think like a computer. However, there are serious downsides to this, especially since it is clear that AI is the future of programming.
**Use AI a lot, but no coding harness.** This is kind of like vibecoing before claude and cursor, just learn how to give the AI assistant context and ask the right questions, but you still end up pasting the code in your own editor, and you do not have the tight integration that is available now. On the bright side, there are lots of these AI assistants that are free for students.
**All in on AI: Claude Code or Codex.** This is how I actually write software today, and it is the most useful, but I worry how much a student would actually learn with a tool this powerful. On the other hand, maybe it is the right place to start because this where they will eventually end up anyway, and we can build really useful things the fastest with this setup.
Of all the options, I am leaning towards option 2 myself (goldilocks option), but let me know what you think.