Since the first days of computers, some have hoped that automation
Would have a major impact on the state of education.
The teaching of school children really seemed like it might be
A killer application for the new technology.
Four million second graders in 100,000 schools
Are studying addition and its complicated rules.
Since all of them are questing for the same scholastic grail
You'd think you could achieve some large economy of scale.
The potential wealth of data makes a data miner drool:
A billion hours weekly of recorded time in school!
When properly anonymized, this information can
Let a teaching program optimize its pedagogic plan.
The advantages of programs are very clear indeed.
Each student gets the one-on-one instruction that they need.
A program's always patient when you're struggling to learn,
And sexual abuse is hardly ever a concern.
And yet, despite this promise, it remains a stubborn fact
That teaching software's failed to have a serious impact.
Schools have lots of different programs doing lots of different things,
But teaching in the classroom is still done by human beings.
As a supposed AI expert, I'd find it hard to name
A profession less amenable to AI's much-hyped game.
You need to grok how kids misunderstand the stuff you've taught.
Evaluation's murder, and politically fraught.
You say this poem is shallow, superfici-al, and bland?
You bet! If I went deeper, I'd be forced to take a stand
On the social role and aims of education as a whole,
And I won't touch that debate with a 20 meter pole.*
* As we say in academia , "That lies outside the scope
Of the poem that you're reading. That's for future work, we hope!"
This is part of the collection Verses for the Information Age by Ernest Davis