It’s a sunny Friday afternoon at Ronald Reagan High School right before the end of the school year.
The students have already left for the day, and teacher Brandt Champion is organizing items from a recent track and field banquet and packing his belongings into boxes. He'll be moving classrooms over the summer.
Champion has been teaching the sciences for nine years, and recently he's been changing the way he teaches with the help of artificial intelligence.
Champion uses an AI tool called Google Notebook LM to create study guides, games to help students review material before exams, and even generate podcasts based on his class notes.
To do this, Champion uploads his classroom presentations and his review packets into the Notebook LM platform. This "trains" the AI model to only answer students' questions and pull material from his notes. No scouring the internet for random sources of information.
Notebook LM generates a podcast on his notes on the skeletal system in less than ten minutes.
The podcast is set up like a conversation between two AI-generated "hosts." They don’t quite pass for people, but they also don’t sound like robots right away, either.
Here's a little bit of it:
Champion says these podcasts have been wildly successful for his students, who report listening to them on the ride to school to review. Around 80% of students have accessed the AI-generated tools.
“In my classroom, I encourage students to use these AI tools as that: as a tool, as a way to help you understand what you need to achieve better," he says. “On final submissions and where you are trying to demonstrate mastery of a skill, it's absolutely zero tolerance of AI-generated work.”
In education, the pendulum on new technology swings mightily back and forth. Just two years ago, educators from K-12 up to the college level were wringing their hands about how they were going to catch students who were cheating using AI. Many were swearing it off entirely at the time.
And that definitely still exists.
But in the last year, more educators have begun to advocate for teaching AI literacy so students can grow up and use the tools responsibly.
“No matter how a teacher feels about AI from like their worldview, we have to teach how to use these responsibly and successfully," Champion says. "We also have to teach students about the cost of AI as well. Every time I’ve introduced an AI tool, that conversation has come up in every single one of my classes.”
Those costs are becoming more known: The data centers used to power AI are taking over parts of rural America, sending utility costs soaring in some cases. Authors, illustrators and creators argue that AI trained using their work is stealing their intellectual property. And experts warn that using AI can hinder a child's critical thinking and reasoning skills.
UWM music professor says college students shouldn't trust sourcing provided by AI
Like in high school, there are also a lot of differing opinions on AI in higher education.
Gillian Rodger is the chair of the school of music at UW-Milwaukee. She pushes back on using words like “thinking” or “creating” when talking about the technology.
“When you use AI, first of all, you can only trust it to make plausible combinations of words, which is all it does. There's no intelligence there. There's no knowing in AI," she says. "You can only trust it to work on things that are already well known. The other problem is that you cannot trust that the sources that it's drawing on are reliable because you don't know what they are.”
Rodger teaches a history of music class as a general education course, but she spends most of her time working with musicians and performers.
To illustrate the potential dangers of AI, Rodger allows students in her general course to use it for one assignment at the start of the semester.
“I ask them to come up with a prompt for an essay, and that they must apply at least five to ten bibliographic citations at the end. And they've got to figure out, how do they ask AI to get those results?" she explains.
Rodger says the essays and citations usually aren’t very good. That’s the point: sourcing and citing works of others is a crucial skill for college and graduate level students.
Ban it or integrate it? Colleges and faculty torn on AI use
AI at the university level is a messy topic right now.
While some agree with Rodger that AI represents a fundamental threat to the humanities, there are other departments and other colleges re-orienting their entire curriculum to implement it.
In March, the Milwaukee School of Engineering announced it would integrate applied AI into all degree programs. At the time, MSOE president Dr. Eric Baumgartner said the school's AI framework “guarantees that every graduate gains the practical skills, ethical grounding and adaptive mindset needed to lead in an era where AI influences every industry."
In other corners, researchers are flying through data analysis they used to hire a dozen graduate students to help sift through.
But Rodger says that what makes AI useful in that kind of number crunching also makes it less helpful for her students.
“Rather than looking at vast quantities of raw data, what we're looking at is previous bodies of scholarship from prior generations of scholars and asking what hasn't been asked. Where are the gaps in this? How else might I think about this?" she says. "And that is why AI fails so spectacularly in the humanities realm.”
She cautions that what works for younger students in middle and high school may not work for university students — who may effectively be in a different generation based on when they started using AI.
The same goes for a technology that works on one corner of campus but doesn’t work in another.
“I'm lucky in that I'm on committees with scientists who've been able to articulate to me why it works for them, and I'm happy for them. I want them to use it," she says "If it's a hammer, I don't want all of us to be nails.”
Do you have a question about education or how schools work in our area? Submit it here to WUWM education reporter Katherine Kokal.
_
