AI in classrooms? Cornell University professor brings back typewriters

At Cornell University, German language instructor Grit Matthias Phelps has begun requiring her students to complete certain assignments on manual typewriters. Laptops, Wi-Fi, translation tools, AI, nothing is allowed. If a sentence goes wrong, it stays wrong—or the page gets redone from scratch. She has made an academic environment that is stripped of frictionless perfection.

Cornell professor is bringing back typewriters in class

Grit’s reasoning is that when students submit work that is 'too perfect', it becomes impossible to tell whether they actually wrote it. In a world where AI can produce grammatically pristine paragraphs in seconds, fluency is no longer proof of learning but a commodity. So she has introduced what she calls “analog assignments,” which forces a return to visible thinking.

The benefits, at least on paper, are clear. Students slow down, make deliberate choices instead of predictive ones and confront the mechanics of language and not outsource them. Errors become part of the process instead of something invisibly corrected. There is also more collaboration; students turn to each other rather than to a search bar.

In other words, the typewriter restores evidence, that education has lost.

Because what’s happening in that classroom is containment. For decades, education has moved toward greater efficiency with spellcheck, grammar tools, online research and collaborative documents. Each step was framed as progress, but now, those same tools can replicate the end product of learning without requiring the process itself. If a student can generate an essay indistinguishable from their own work, what exactly are we assessing?

Grit’s solution is to remove the ambiguity by removing the technology. It works, in a narrow sense. A typewritten page is undeniably human. The hesitations, the inconsistencies, the small mechanical struggles will all leave a trace.

It exposes how quickly the meaning of writing is changing. It forces the educators to confront whether they are teaching a skill, a process, or just grading an output. And it reveals that the line between actual human effort and machine assistance is dissolving faster than our ability to respond.

The image of a student striking keys on a stubborn, ink-ribboned machine might seem insignificant. But it is also a warning. When the only way to guarantee original thought is to unplug entirely, the problem is not the assignment, but the world we’ve built around it.

For more updates, join/follow our WhatsApp, Telegram and YouTube channels.