Short Take On One Aspect Of AI & Schools + New Wonkyfolk

Jed Wallace and I have a new Wonkyfolk out this morning. We cover a few issues, but in particular what Ron DeSantis and the teachers unions have in common – love of theater and they sure don’t like to take yes for an answer. Plus NAEP and civics, The Economist, more…

You can also read or watch via this link. Or listen wherever you get your podcasts.

This post is not written by AI, that’s been done enough. But it’s about AI. An observation and a question for you.

A few weeks ago John Bailey posted a funny riff on the Good Will Hunting library card scene, but with an education flavor:

There is lot of speculation about what kinds of jobs might be most disrupted by AI. People have realized the issue is no longer just truck drivers and retail jobs, the traditional targets of automation. AI is a whole new ballgame. AI can do entry level journalism, ad copy, and many other tasks that seemed at least somewhat insulated from disruptive technologies. It is primed to revolutionize tutoring and likely parts of the teaching job. That has a lot of people freaked out.

In our little part of the world even now I suspect that AI can do 80% of what an early career analyst or writer in this sector does. But that 20% matters. A lot. It’s still a funny joke, but in the example above the characterization of Yong Zhao and to some extent Darling-Hammond is off. That’s where the 20% comes in. You can’t just plug and play, you have to know something about content to learn new things.

The other day we looked at history and civics and how part of our history and civics problem is we don’t teach civics and history content. Content gets supplanted by the allure of teaching “skills” and the education field is always a sucker for that play.

This is where it might really matter. If AI can do 80% but the last 20 requires domain knowledge/subject matter expertise then that’s one way AI could increase gaps between education haves and have nots. Sure, AI will get better – and quickly – and in a few years Bailey’s script would likely be more refined. But although the ‘we don’t need to teach content with this great new technology’ crowd is making themselves heard again, it might be that we need content more than ever.

How do you like them apples?

Like a lot of organizations we’re doing work around AI at Bellwether. One thing I’m interested in the education space is the most interesting unsanctioned uses around schools. So there are lots of applications prek-12 emerging. But what uses are students figuring out themselves. In particular examples of younger (elementary and middle school) students using AI on their own? Creative informal student led applications? Other things that are less common? Please send me any ideas, anonymous or blinded is fine, to eduwonk AT