Feedback For SNL's Generative AI In The Classroom Skit
Saturday Night Live aired a skit with a teacher using generative AI to teach her students.
I do not recommend playing this video for students because of some sensitive subject matter. I wrote a post featuring resources teachers can use to talk with students about generative AI.
Resources For Talking To High School Students About AI
Teachers may be frustrated if students use generative “AI” to generate assignment submissions. They may take offense if students don’t know about the environmental racism, theft, and worker exploitation inherent to generative “AI.”
As a former classroom teacher who writes about pedagogy and generative AI, I will put on my critical thinking cap to give the skit some feedback. The show has been a boon to my writing considering I used legendary SNL character Irwin Mainway to write about generative AI in November. My feedback has two parts: a running record of observations with timestamps and some big-picture thoughts.
Before sharing observations, here are the characters in the scene:
Ms Rotman, the teacher, is played by Ego Nwodim.
A student played by Marcello Hernandez. I will refer to him and all students by the first name of the actors playing them. No student characters are named in the skit.
A student played by Michael Longfellow.
A student played by Sarah Sherman.
A student played by Andrew Dismukes.
A student played by Jane Wickline.
AI-generated podcast hosts played by Timothée Chalamet and Bowen Yang. Their names change throughout the skit.
Skit Observations
0:39 - Ms. Rotman says, “So the school has invested in a new AI program that takes your textbooks and turns them into an educational podcast.” This must refer to Google’s NotebookLM, which has an audio summary feature that generates an audio “podcast” from content in a file. The difference in the skit is that the “podcast” is a video. I tried NotebookLM and wrote about my experience.
Students Deserve Human-Voiced Audio
Amidst the public dislike of AI and its frequent failure, the AI in education hype continues unabated. Last week, Forbes published an essay arguing Tesla’s humans disguised as robots “have the potential to integrate into various aspects of daily life, including educational settings, potentially revolutio…
1:28 - Yang: I’ve been thinking, what should I put in my new amazing project, the Constitution?
Chalamet: Oh man, great question.
This snippet of dialogue quickly catches what I have noticed when people share their NotebookLM-generated podcast files: Overstated positivity. Whatever the topic, the podcast hosts speak like it is the most exciting thing in the world. It feels like something Larry David said about Ted Danson on Curb Your Enthusiasm, “Everything’s heaven with him. The piece of gum he had, ‘Oh, this is heaven.’ Had to taste a chocolate bar. ‘I’m in heaven.’ The parking space is ‘heaven.’ It’s all ‘heaven.’
1:33 - Chalamet says, “Maybe something about the housing of soldiers in civilian homes during peacetime.” As a former U.S. history teacher, I know the 3rd Amendment prohibits quartering soldiers during peacetime. The skit captures a weakness of generative AI here: It has word form but lacks word meaning. The language is correct, but because generative AI does not understand what it means, it is presented without enough context to learn.
2:36 - Ms Rotman says, “See, it’s so fun it doesn’t even feel like you’re learning.” This condescension captures gimmicky generative AI, such as historical figure chatbots, marketed to kids.
2:38 - Andrew says, “I don’t think we are learning.” This is a possible outcome of using a technology not well-matched to applications where the accuracy of the content matters for learning.
3:13 - Chalamet raises his hand, showing six fingers. Yang then does the same, referring to generative AI’s well-known inability to depict human hands.
3:25 - Andrew: Why would a girl propose?
Sarah: Trish sounds butch AF. (audience laughs)
The low point of the skit. Is a woman proposing to a man funny? What a retrograde joke from the SNL writers! Further, generative AI would probably generate the dominant outcome in its dataset: A man proposing to a woman. Generative AI will most likely generate the mean outcome, not Lorelai proposing to Luke.
3:36 - Yang says, “Trish! Trish! Over here! I saved you a seat at brunch!”
This calls back one of my favorite SNL skits. Its title is “Ladies Who Lunch,” but come on, they are drinking mimosas - it’s brunch.
3:48 - Ms. Rotman says, “Women named ‘Trish’ can look like anything.” That is not how generative AI image generators have performed. As The Washington Post documented, “AI image generators like Stable Diffusion and DALL-E amplify bias in gender and race, despite efforts to detoxify the data fueling these results.”
4:43 - Andrew says, “You know what, Ms Rotman? Let’s just read our textbooks. This is making me realize we have something better than AI. Our brains.” Andrew encapsulates a point Benjamin Riley recently made.
“Reputable websites and textbooks are more reliable than AI. Maybe teachers and students should just use those resources in the first place?” - Benjamin Riley, December 4, 2024.
4:59 - Marcello says, “This is what AI is for,” and presses a button on a remote control to have the two AI-generated podcast hosts dance. Marcello does this without entering a prompt (not how it works), but at least generative AI is not promoted as a teacher or tutor. Still, using AI to generate silly content does not address inherent harms such as racism, misogyny, theft, worker exploitation, environmental degradation, and encouraging self-harm.
Big Picture Feedback
Here is some feedback for the SNL writers across two domains: generative AI insights and humor.
Generative AI Insights
What the skit did well to explain generative AI includes:
Captured the overly excited, inauthentic tone of NotebookLM-generated podcasts.
Showed that generative AI is good at word form, not word meaning.
Captured how generative AI is marketed to kids.
Expressed reasonable doubt about whether generative AI aids in learning.
Demonstrated the generative AI human hand issue.
What the skit needs to improve to explain generative AI:
Generative AI usually generates content that looks like the content in its dataset. It is unlikely it would generate content where a woman proposes to a man unless specifically prompted to.
“A woman named Trish can look like anything” does not convey how image generators often generate biased, sexualized images of women.
A minor critique: Generative AI works with prompts, not pressing a button on a remote control.
Humor
The skit is more amusing than laugh-out-loud funny. The highlight was the callback to Trish (such a well-written, exquisitely acted skit). The worst joke was about “Trish” proposing to a man. How did that make it to air in 2025?
Let’s Talk
What do you think? What do you think about how SNL commented on generative AI and education? Comment below or connect with me on BlueSky: tommullaney.bsky.social.
Does your school or district need a tech-forward educator who recognizes the harms and limits of generative AI? I would love to work with you. Reach out on BlueSky, email mistermullaney@gmail.com, or check out my professional development offerings.
AI Disclosure:
I wrote this post without the use of any generative AI. That means:
I developed the idea for the post without using generative AI.
I wrote an outline for this post without the assistance of generative AI.
I wrote the post using the outline without the use of generative AI.
I edited this post without the assistance of any generative AI. I used Grammarly to assist in editing the post. I have Grammarly GO turned off.
You're right on every point--especially how AI does not challenge biases but reinforces them. This is such an overlooked premise that's maybe too heavy for even SNL to handle on live TV.
I loved the skit, and the punch line about reading actual textbooks!