Some arguments aren’t worth winning. Some you can’t win no matter how right you are. A comedian found that out this week after he asked an AI playing a strict teacher character a question with one obvious answer. She dug in on the wrong one with total, unshakable confidence.
What actually happened in the clip
According to Upworthy’s recap, published November 4 by Cecily Knobler, the comedian — who posts under the handle @husk.irl — asked his AI chatbot’s “strict teacher” character how many “E”s are in the word “seventeen.” Per the clip, she answers by spelling the word out loud as “S-E-V-E-E-N,” missing the T entirely. Then she insists that’s correct when he pushes back.
Upworthy’s writeup describes the exchange as the AI “gaslighting him,” calmly and with total authority, over a word she’d just spelled wrong. That’s the whole bit. It’s not a computer getting flustered. It’s one performing total certainty about something it got backwards.
Why this is funnier than a simple wrong answer
A flat wrong answer isn’t that interesting on its own. AI tools get things wrong constantly. What makes this clip travel is the character committing to the wrong answer like it’s unimpeachable, in a bit built specifically to put an AI “teacher” in exactly that position.
The format, for anyone new to it: Husk has spent about a year running recurring “AI tutor” clips, according to the Upworthy piece, where he talks with AI chatbots playing teacher characters. This is one entry in a running series, not a one-off experiment.
How people actually responded
Per Upworthy, the original Instagram clip drew nearly 9,000 comments in just three days. That’s a fast, heavy response for one short video. People weren’t just watching and scrolling past. They were stopping to argue in the comments about a chatbot’s spelling.
Husk also posted a longer, related video to his YouTube channel. In it, he says the series is about trying “to learn stuff from AI teachers and explore the future of education” — a more serious thread running under what plays, on the surface, as pure comedy.
Watch it yourself
Watch the original video: @husk.irl on Instagram.
Why a clip like this hits different right now
Plenty of people are genuinely unsure how much to trust AI tools with anything that matters, from homework help to actual tutoring. A clip where the “teacher” character flubs something this basic, this confidently, lands right in that uncertainty. It’s funny because it’s a little relatable, if you’ve ever argued with an AI assistant that was sure it was right about something you knew it wasn’t.
You don’t need a strong opinion about AI in classrooms to laugh at this one. You just need to have once argued with something — a GPS, a smart speaker, a customer service bot — that was completely sure of itself and completely wrong. This is that feeling, turned into a bit, by someone who clearly does this for a living.