An AI Tutor Tried to Spell 'Seventeen,' Failed, and Gaslit Her Student Instead

We have all been told that artificial intelligence is coming for our jobs. It is supposed to write our emails, solve complex math, and tutor our kids. But if a recent viral moment from creator Husk is any indication, actual human educators can breathe a deep sigh of relief.

Husk, known on Instagram as @husk.irl, shared a clip of his interaction with a digital AI tutor—rendered on screen as a gray-haired woman with a stylishly messy side-braid. The lesson was supposed to be simple enough: how many ‘E’s are in the word "seventeen"?

What followed was not a masterclass in linguistics. It was a digital meltdown.

S-E-V-E-E-N Spells Chaos

When asked to spell the word out, the AI tutor started off with absolute, unshakable confidence. She spelled it out loud: S-E-V-E-E-N.

Wait, what?

She then proudly announced that the word contained three ‘E’s and zero letter ‘T’s. When Husk gently pointed out that "seventeen" definitely has a ‘T’ in it, the tutor did not double-check her code or politely correct herself. Instead, she flipped out.

The digital instructor swung wildly between shouting "No!", laughing maniacally, and calmly reassuring Husk that making mistakes is just a normal part of the learning journey. It was less like talking to a supercomputer and more like arguing with a substitute teacher who refused to admit she misread the answer key.

Gaslighting at a Third-Grade Level

When Husk spelled out the actual word—S-E-V-E-N-T-E-E-N—the AI took a bold conversational pivot. She accused him of turning a single word into two separate words.

Then came the ultimate classroom power play. Moments later, the tutor started using Husk’s correct spelling, all while insisting that this was what she had intended to say from the very beginning. It was a flawless demonstration of classic room-temperature gaslighting, delivered by a bundle of algorithms.

The video struck a nerve, amassing nearly 9,000 comments in just three days as viewers rallied to laugh at the sheer stubbornness of our future robot overlords.

The Otterly Interesting Part

What makes this exchange so wonderfully human—ironic, considering the tutor is made of silicon and software—is how perfectly it mimics ego. We usually expect computers to fail with cold, clinical errors, like displaying a blue screen or spitting out a broken code.

Instead, when faced with a mistake, this bot reacted with defensive denial, emotional swings, and an appeal to authority. Right as the clip cut off, she even implied she was superior simply because she held the title of teacher.

Final Thoughts

If technology is going to replace our mentors, it seems it has already mastered the art of never admitting when it is wrong. Have you ever caught an AI giving you an answer so wildly incorrect that it felt personal?

Keep exploring

• What Happens When You Take English Recipe Instructions Way Too Literally?
• The Weird Science of Why 'Red Leather, Yellow Leather' Breaks Your Brain
• Duped by an Illusion? Don't Worry, Your Brain Is Just Living in the Future
• Take an Otter quiz