The Robot Ate My Homework

The Robot Ate My Homework: The History of AI in Classrooms

If you traveled back to 1900 and asked a French artist what school would look like today, they wouldn't have described iPads or Zoom. They would have handed you a postcard titled “En L’An 2000” (In the Year 2000), depicting a teacher feeding dusty textbooks into a wood-chipper connected to wires that dangled from students' heads. The idea was simple: grind the knowledge, pipe it directly into the brain.

While we (thankfully) avoided the cranial book-grinder, the journey of artificial intelligence in education has been no less bizarre.

The Wooden Teacher Long before ChatGPT was writing C-minus history essays, there was Leachim. Built in 1974 by Michael Freeman for a Bronx elementary school, Leachim was a six-foot-tall robot with a wooden body, a distinct robotic drone, and a surprisingly good memory. It knew every student’s name, their parents' names, and even their hobbies. If a student struggled with math, Leachim was patient. If they did well, it might ask them about their pet dog.

It was a technological marvel, but it was also terrifyingly heavy and looked like a grandfather clock had a baby with a vending machine. Leachim was eventually stolen during a transport run, disappearing into the annals of history—perhaps the only teacher to ever play hooky permanently.

The "Eliza" Effect In the 1960s, MIT professor Joseph Weizenbaum created ELIZA, a chatbot that parodied a psychotherapist. It mostly just repeated your words back to you as questions ("I am sad." -> "Why do you say you are sad?"). Despite being about as intelligent as a toaster, students and staff would spend hours pouring their hearts out to it, convinced it genuinely cared. This phenomenon—projecting human intelligence onto dumb code—became known as the "ELIZA Effect." It explains why we still say "Please" to Alexa, just in case the robot uprising starts tomorrow.

The Modern "Copy-Paste" Era Fast forward to today, and the advent of Large Language Models has turned classrooms into a digital Wild West. We have entered the golden age of the "clumsy cheater." Teachers now regularly receive essays that begin with the dead giveaway: "As an AI language model, I do not have personal opinions, but..."

Yet, for every student trying to outsource their homework, there’s a quirkier, more creative use case emerging. We see "Socratic" bots that refuse to give answers, instead annoying students with endless questions until they figure it out themselves (a digital simulacrum of your strictest math teacher). We have AI tools that can generate "funny pictures of chickens" to explain complex economic theories, proving that humor remains the best hook for learning.

The Future? We may never get the book-grinding machine of 1900, but we are drifting toward something stranger: personalized AI tutors that know us better than we know ourselves. They won't be made of wood, and they (probably) won't get stolen in the Bronx. But as we let algorithms guide our learning, we must remember: the goal isn't just to download information. It's to keep our own "wetware" (our brains) working hard enough that we don't accidentally become the robots ourselves.