AI Has Changed the Wrong Part of Education
Every generation gets a technology that makes the previous generation nervous, and looks like it might wipe them out entirely.
Do you remember any of these? Because (kids) they all actually happened.
Calculators were going to destroy mathematics
Wikipedia was going to destroy research
Google was going to destroy memory
So when it’s AI’s turn we should not be too surprised that although “destroy” has been replaced with “We’re cooked” but the mentality is still the same.
It’s the idea that “won’t you think of the children”. It doesn’t make sense, and it’s just here to stop progress.
Students use the technology available to them
They always have, they always will do.
It might be books. It might be tech. Who really cares?
Because it’s not interesting to ask whether students should use AI. The real question is why so much of our educational system can apparently be completed by an AI in the first place.
That sounds harsh, but hear me out.
If a machine can complete an assignment from start to finish with minimal guidance, what exactly was that assignment measuring?
Was it measuring understanding?
Or was it measuring whether someone could follow a process?
Software developers are nodding their heads and appreciate that this is really the same problem as they have faced for years.
The Same Idea Repeating
As software developers we’ve seen this before.
For years companies interviewed engineers by asking them trivia questions. Memorize enough facts and you passed. Forget one detail and you failed.
The problem wasn’t the candidates.
The problem was that the process rewarded memorization instead of competence.
That’s as familiar to educators as it is to software developers.
Our world used to mean producing an answer was expensive. Today producing an answer is effectively free, shifting our value (and values)
Knowing the answer matters less than understanding whether the answer is correct.
So I’ll put it to you. Do you understand all of the code pushed in your name these days? Or have you outsourced the whole understanding to Claude.
AI Danger
AI can generate code. It can generate tests. It can generate documentation. What it can’t reliably do is understand the broader context of a business problem, challenge requirements, or take responsibility when something goes wrong.
Someone still needs to review the output.
Someone still needs to think.
The skill that matters isn’t producing 500 words on a topic in an hour. The skill is understanding whether those 500 words are any good.
Unfortunately, that’s harder to measure, so most companies don’t bother.
Something to think about.
About The Author
Professional Software Developer “The Secret Developer” can be found on Twitter @TheSDeveloper.
The Secret Developer stopped thinking decades ago.