The curious case of engineered errors

The future of education, I recently discovered, arrived not in a government white paper or a Silicon Valley keynote, but in the bedroom of my twelve-year-old nephew. He had been assigned an essay, a perfectly ordinary school task, designed to cultivate clarity, argument, and intellectual temperamant. He opened his laptop, summoned a chatbot, and issued an instruction that was at once comic and chilling: “Write the essay, but make a few small mistakes so my teacher doesn’t realise it wasn’t written by me.”

There it was. No sense of plagiarism in the old-fashioned sense, but a performance management to get the right score. The machine was to think for him as well as simulate adolescent fallibility. It had to sprinkle in minor blotches, the academic equivalent of cosmetic freckles, so that authenticity might appear intact.

We might have dismissed this as youthful cunning were it not for the recent episode involving a platform brazenly named Einstein AI. Unlike the gentler educational “assistants” currently marketed by major technology companies, Einstein AI did not confine itself to offering suggestions or summarising lecture notes. Einstein AI promised to log into students’ virtual learning environments and complete their assignments “while they slept.” It would even submit the work from the student’s own account “just like you would,” having digested lectures and course material in advance. A perfect academic doppelgänger. No late nights. No panic. No existential wrestling with footnotes.

Predictably, outrage followed leading to the name vanishing from public view almost as quickly as it had appeared. But its brief existence exposed a trajectory that many prefer not to confront. If a chatbot can generate a competent essay in seconds, what remains of the educational process beyond editing and concealment?

My nephew, without knowing it, had grasped the logic perfectly. If the machine produces flawless prose, suspicion arises. Therefore, imperfection must be engineered. Interestingly, we have reached the point where human weakness must be algorithmically manufactured in order to preserve the illusion of human effort.

The defenders of these systems insist that they are merely tools. And of course, every intellectual age has had its tools. The printing press multiplied access to knowledge. The calculator relieved us of tedious arithmetic. The Internet collapsed distance. But tools traditionally extended human capacity and did not smother the human agent altogether.

The image of the ship sailing across the globe collecting knowledge came to my mind. It was Francis Bacon once imagined knowledge as a disciplined voyage, a collective enterprise in which inquiry, experiment, and observation advanced human understanding with ships venturing outward, gathering discoveries, returning enriched. Knowledge required movement, engagement, and daring. What would Bacon make of a system in which the voyage is cancelled and the cargo delivered by a drone?

I remember, early on in school, my teacher introduced us to Cardinal Newman’s The Idea of a University, where he argues that education is the cultivation of the intellect, the training of judgment through sustained engagement. The essay, in that tradition, is not merely a product to be evaluated but a process where drafting, faltering, revising are not inconveniences but the very substance of learning. Remove the process, and you may still produce a text but not produce an educated mind. I learnt this when still in junior school where writing the essay was our daily task. Confronting the mute page with a pen in hand and no downloading or copy-pasting

I realized, my nephew’s strategy reveals the cultural shift beneath the technology. He is not rebelling against education; he is adapting to an environment in which outcomes are prized above inward growth. If the grade is what counts, and the machine can secure the grade, then so be it. A few inserted mistakes become camouflage, a proof not of stupidity but of strategy.

We should resist the temptation to treat this as harmless ingenuity. When systems are built that can replicate the visible markers of learning, institutions risk mistaking simulation for substance. The question raised by Einstein AI is not whether such platforms should exist but whether universities are prepared to confront what they imply. If assignments can be delegated entirely, assessment itself loses meaning.

There is something faintly comic about instructing a machine to “add mistakes.” Yet the humour carries an edge. We are teaching a generation that authenticity is a surface effect to be programmed. We are encouraging them to treat thinking as a service that can be subtly disguised. The only unmistakably human trace left in the essay is the deliberately planted error.

shelleywalia@gmail.com