Learning to use AI
Kids use AI for
their schoolwork. Plenty of them use it, not as an assistant, but as the entity
that does all the work. That is a problem obviously.
This blog is not
on a solution for that problem (None exists. Not yet anyway). Instead, this
blog is based on a post by college student Maximilian Milovidov on a course called Writing
AI. What’s unique about it?
“(It)
might be the only one on campus where artificial intelligence was not
prohibited but, rather, required.”
The spirit of this
course is an interesting experiment:
“What
if we taught students to use AI critically, rather than insisting they ignore
it or assume they're using it to cheat?”
AI, after all, he
says, is here to stay. You can’t wish it away any more than our ancestors could
wish away the printing press.
Here’s how the
course works. Students have to bring their own ideas and outlines to the class.
“We
fed drafts into a chatbot while documenting its suggestions and then explaining
why we accepted or rejected them.”
The idea is to
learn to use AI as a “friend”, as one of his professors put it.
“We
began with our own sparks of inspiration, argument and thought, but learned to
prompt chatbots to expose gaps in reasoning or find unseen connections.”
The key points
here are (1) come up with the proposal oneself, (2)
use the AI to identify flaws and misses, and (3) evaluate whether
the AI’s feedback is right and relevant.
This course
teaches the student how to use (language) AI. If done in that spirit and
manner:
“Students
using moderate AI assistance during lectures outperformed both the students
using fully automated help and those using minimal support.”
An interesting
course, no doubt. I do see three limitations. (1) Identifying
when an AI is wrong is a chicken and egg problem. To spot an error, you need to
be knowledgeable on the topic. But you are using the AI while you are still
learning! (2) The ability to question, to doubt, to cross-check
is an adult skill. Sure, college kids are almost adults, so they could do it
with practice and training. So this technique probably can’t be used for school
kids. (3) The course in question is all language-based content
with subjectivity, but what about the objective/absolute worlds of math and
science?
Then again, this is a start. The first step in a journey, where we will evolve new techniques and change course to live and work in the world alongside AI.
Comments
Post a Comment