Hi everyone, I'd like to share some thoughts I have on AI in student life.
Nowadays, the internet already gives us access to great material to learn new topics: videos, articles, you name it. So what we probably need next is feedback on the tasks we do independently to follow up on that material, or someone to explain the parts we don't understand. That's where I think AI helps: we can start asking questions, and ask for personalized explanations based on our own learning preferences, analogies, examples, and short exercises.
The way I like to self-study is to watch videos first, since I find that a fast way to get introduced to a new topic. Then I write down the points I don't understand and ask AI about them, go back to the video, or move on to textbooks or papers. After that, I try some practical exercises, and if the main source doesn't have any, I ask AI to generate one. To be honest, I don't feel the exercises from a first prompt are usually good, so you probably need to iterate to actually cover the skills you're looking for.
I'm still reflecting on this process. It isn't perfect, but I consider it a solid way to use AI as a teacher. If you read this and use AI to learn in a different way, don't hesitate to send an email. I'm always looking for new tips.
Another point of view, this one from a conversation with one of my mentors during my SURF internship, is about how to use AI. The ideas below aren't mine; I'll lay them out first and then add my own take. The starting point is to define your goal: what do you actually want? If you want a product, a fast solution is to use AI in an autonomous way, for example with Claude Code, to generate something cool. But if your goal is to learn, you probably need to avoid letting AI work autonomously for you.
That tip stuck with me, and it made me reflect on how I use AI. For students pursuing a PhD, for instance, the goal is usually to learn, so it's worth using AI the way I described earlier. For something like building a personal website, though, letting AI move fast and autonomously makes more sense.
A related point, from a different angle this time: as a teaching assistant, seeing young first-year undergrads makes it clear that education has changed, and that teaching students how to use AI properly matters. I only started using AI in my fourth year, once I already had a basis in programming and had formed some criteria for making technical decisions. As teachers, we can't avoid or forbid students from using AI; I consider it part of the new reality. Our challenge is figuring out how to teach students to use it well: defining the new skills this situation calls for, writing better prompts, designing an architecture or sketching ideas on a notebook or whiteboard, being the designer (which I like), and then letting AI help turn that idea into reality. Creativity and generating new ideas is what humans bring to the table, and I'm still exploring new approaches to using AI and to teaching others to use it. If you have ideas of your own, I'd love to chat.