[ExI] Maths Teachers Career Prospects

[BillK]

On Tue, 20 Oct 2026 at 04:55, spike jones via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

> I was reading an online group chat among students I once coached in
> competition math, who are now seniors and grad students.  Most of them
> appear convinced it was a waste of time and resources to study math.  I
> don’t post there, only read, so I couldn’t encourage them.
>
>
>
> A comment really struck me.  One of my former students is working toward
> becoming a math teacher.  He opined that if it becomes meaningless to study
> math, it is also meaningless to become a math teacher.  Reasoning: AI can
> teach it better, faster, cheaper, it can adapt to the students
> shortcomings, focus on teaching at the weak points, fill in the gaps, etc.
> He is right.  Dwayne, If I can get a message to him, any suggestions?
> Anyone else?
>
>  spike
> _______________________________________________
>


I discussed this with DeepSeek AI.
Finally, I asked DeepSeek to draft a letter to the student.
I think this will be helpful.
BillK


DeepSeek V4 AI:

*Subject: Your question about math teaching — and why I think you’re asking
the right thing*

Dear [Name],

I’ve been thinking about what you said—that if AI can teach math better,
faster, and cheaper, then studying math becomes meaningless, and so does
becoming a math teacher. I want to take that seriously, because it’s not a
silly worry. It’s the kind of question a thoughtful future teacher *should*
ask.

You’re right about a lot of it. AI is already very good at the diagnostic
and procedural parts of teaching: spotting gaps, generating practice,
giving instant feedback, adapting to a student’s weak points. If teaching
math were only that—transmitting procedures and correcting errors—then yes,
a machine would increasingly do that part well.

But I’d push back on the conclusion, because it assumes teaching is *only*
that. It isn’t.

The evidence here is quite striking, and a lot of it comes from American
classrooms. The National Council of Teachers of Mathematics, the largest
math education organization in the US, states plainly that students “will
continue to need teachers’ mathematical, pedagogical, and relational
expertise,” and that contrary to popular opinion, integrating AI well will
require *“even deeper knowledge of math instruction and assessment—math
teachers with more experience, not less.”* A two-year randomized study in
18 Tennessee middle schools makes the same point from a different angle.
Students were given access to an AI math tutor called Khanmigo, designed to
guide rather than give answers. Almost everyone tried it, but most used it
rarely, sent it off-topic messages, or tried to get it to hand over the
answers. The learning gains were real but small—and comparable to what
students got from ordinary Khan Academy practice *without* the AI. The
researchers’ conclusion was not that AI is useless, but that *access to AI
is not the same as learning with AI*; students need active encouragement
and human oversight to use these tools well.
<https://t3n.de/news/ki-als-nachhilfelehrer-langzeitstudie-beweist-warum-echtes-lernen-weiterhin-menschliche-betreuung-braucht-1760578/#1>

A math teacher’s real work is not just getting a student through the next
problem set. It’s helping them learn how to think: how to ask a good
question, how to sit with confusion, how to build an argument, how to
notice when an answer is elegant or when a proof has a hole. It’s seeing
the student who has gone quiet—not because they don’t understand the
algebra, but because they’ve lost confidence—and knowing how to bring them
back. AI can support that work. It cannot *be* that work.

And studying math is not becoming meaningless. If anything, it’s becoming
more important. AI is built from mathematics—linear algebra, probability,
statistics, optimization. The people who can reason clearly, model
problems, and check machine output will matter more, not less. The goal of
studying math was never just to compute. It was to learn a disciplined way
of thinking—and that is still one of the most valuable things a person can
learn.

I also want you to know that the career you’re considering is not
shrinking. In the US this year, *one in eight teaching positions is either
vacant or filled by someone not fully certified*—and math is among the
hardest-hit subjects, with *40 states reporting shortages*. That shortage
is not a temporary blip; it has been climbing for three years, driven
largely by mid-career teachers leaving. Districts are responding with real
money: some states pay math and science teachers up to *$15,000 extra per
year*, and many districts offer signing bonuses and math-specific
stipends. When
analysts map AI exposure onto US occupations, secondary school teachers do
appear in a relatively “high” exposure band—but the source explicitly
cautions that this is a *relative rank, not a probability of automation*,
and it does not distinguish between tasks that AI will *replace* and tasks
it will *augment*. The tasks most exposed are procedural—grading,
record-keeping, routine preparation. The core of the job, the part that
happens in the room with students, is not what the index is measuring.
<https://www.linkedin.com/posts/k12-careers_the-us-teacher-shortage-in-2026-which-states-activity-7490803329703473153-6dfk>
<https://aichanging.work/en/occupation/secondary-school-teachers?all=1#release-native-tasks>

So I don’t think the career is disappearing. I think it’s changing—from
being the main source of information and practice to being the architect of
the learning, the coach of the thinking, and the human presence that makes
struggle feel worthwhile. In a country with a serious math teacher
shortage, that is not a diminished job. It may be a more essential one.

You are asking exactly the right question at exactly the right time. Don’t
let the fear of being replaced stop you from becoming the kind of teacher
who uses every tool available—including AI—while doing the things only a
human teacher can do.

I’d be glad to talk about this more whenever you want.

--------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20261020/85abad0e/attachment-0001.htm>