Mamdani bans AI in NYC schools(nytimes.com) |
Mamdani bans AI in NYC schools(nytimes.com) |
non-paywall: https://www.nytimes.com/2026/09/01/nyregion/ai-ban-schools-n...
The ban goes through eighth grade, and in high school apparently there's a specific AI literacy class. It's unclear from the article if that's the only AI in high school or it's just unregulated.
My opinion: it's a good policy and frankly just common sense.
- Students using AI risk offloading their cognition onto the AI and thereby skipping the actual learning process.
- Teachers using AI likewise do a disservice to the students.
In the middle of all that, there may actually be a version of AI that can help people learn without incurring the major downsides, but I've not seen any evidence of that in the current AI systems, to the point that an outright ban in the school setting makes sense. I'm heartened to see more education organizations think similarly [1].
[1]: https://www.law.berkeley.edu/academics/registrar/academic-ru...
It’s just refusing to think.
Still, not shocking that the DSA stronghold that elected someone like Mamdani is going to be passionate against AI.
Will be interesting to watch outcomes diverge over the next few years. If you’re gonna ban stuff, do it at local levels like this so we can see differences play out across many locations/districts.
Even when they teach it, screen time will be limited to 45 minutes per week.
It's not about 'passion against AI', it's a year waiting period. There isn't much evidence that AI improves student outcomes, and there's more cases showing harm, so the responsible approach is to wait for more information.
Also, they've banned phone use for quite a while now, which I think most people agree is correct in a learning env.
Grew up in Greece and was shocked that calculators are used in high school math.
An interesting note: "And teachers will not be able to ask A.I. to grade assignments." This prevents teachers from using it as well, which is an interesting angle.
skittish
"skiddish" isn't a word, the only thing I can think of is leaving a 'skid mark' in one's underwear...
The question is "do you want to teach the tool and its use or do you want to teach thinking manually". The west is under the impression that if you don't teach the tool then people will be unemployable.
edit: not even with AI. Because I know there are people here who believe that AI will turn into a superintelligence so advanced that it will literally rewrite the laws of reality like the Eschaton from Charles Stross' Singularity Sky but that also isn't possible.
Moreover the pure egalitarian and voluntarist society of Star Trek isn't plausible. Humans aren't simply going to "evolve beyond our base desires," as Picard once put it. Even if we did have Star Trek style replicators, free energy, teleportation and FTL, and everyone had everything they could possibly want freely provided to them without cost or obligation, people would still want to have more than their neighbor to dominate and control, to be king of the mountain. They would still hate and kill.
You can't educate people out of human nature, and every manner of abomination can and will be justified by someone's ethics or cultural values.
You will still have hate. People will still kill. But one could look at Star Trek and imagine it's plausible to live in a society that doesn't tolerate school shootings or the systemic breakdowns that led to that.
We already have that at least, it's basically anywhere but the United States.
Unfortunately if Columbine and Sandy Hook didn't change things for the better I think the US is a lost cause in that regard.
No. There's proven research that shows reading a book or writing with a pen or pencil is much more stimulating and beneficial compared to looking at a screen or tapping/typing on it.
My point was more that, I think it's incorrect to assert that it's necessary for kids to directly use a computer in order to effectively learn about them. We have tons of concrete evidence of kids learning to be excellent at programming and understanding computers without directly using one. It happened with the kids I described.
It's cool and fiction is how we stress-test ideas before they become policy.
Maybe we should have them live in 15th century lifestyle towns! Wouldn't want them having any contact with this awful world!
Stress-testing only works in reality, where the stresses are also real, or even in a simulation of reality.
Star Trek is neither, it's just fantasy. It doesn't test anything.
I can't help but think of this Chesterton Quote:
"It is, on the contrary, a thing analogous to writing one's own love-letters or blowing one's own nose. These things we want a man to do for himself, even if he does them badly. I am not here arguing the truth of any of these conceptions; I know that some moderns are asking to have their wives chosen by scientists, and they may soon be asking, for all I know, to have their noses blown by nurses."
I can see even a well-meaning student stuck in this aspirational cycle of “I’ll do it myself next time, but right now I just want to catch up.”
Until that will chance the any and all shortcuts will be used, because the goal of the kid is to:
* get grades high enough to keep parents happy * do fun stuff
The motivation to get knowledge for something other than your own hobbies/curiosity only comes when you need to start paying the bills and get a job
I see this everywhere, with people seemingly losing the ability to use the tiniest amount of intellect or mental effort, if only AI can give something that vaguely looks like it might be the answer. Even if the output is great and AI is a great force for the better, and we will all share in its glorious outputs - we’ll degenerate into a lazy, stupid mass unable to reliably count our own fingers.
And that’s a future to be afraid of.
I burned 100s of hours into my 400 level classes, some of those hours were completely wasted on things that I could never figure out how to get them to work.
It’s terrible.
Big regret.
Edit: For reference, the regret came when I tried to do some RF related projects on the side. Those projects would likely have gone somewhere if I put some effort into even the basic calculus from high school.
When I was young I attended a karate dojo. One interaction that always stayed with me was when someone asked why the guy with a black belt could do the katas differently.
The sensei answered that when learning, your task is to learn to do the katas perfectly. As a black belt, you have amply demonstrated that, so now your task is to be creative and build on the foundations.
I liked that answer, it applies to everything. Learn the basics before you are allowed to question them. After that, you must question them.
giving ai to children without limits is like giving them leg-braces or wheelchairs; they'll get used to it and never learn to walk or run by themselves much less karate... they'll be cripples from the start.
This bans it from all teaching tools and materials as well.
Yeah, a bad reason.
https://www.abc.net.au/news/2015-08-10/are-calculators-affec...
> In a 1997 review of the extent to which calculators were being used in schools, numerous studies were examined that indicated the use of calculators at primary levels had no detrimental effects on students' arithmetic abilities.
> Unfortunately, the research indicated that calculators were still being used for trivial things like checking answers and were making little difference in mathematics education.
> Although teachers indicated their support for the use of calculators across all levels of primary school, there was little evidence that these ideas were being taken up and implemented.
> Parental disapproval of the use of calculators was cited as a possible cause of the limited take-up.
That's the part this study seems to handwave. We don't need a full calculator ban, but very controlled usage is necessary.
because you chose computer science/SWE instead of Computer Engineering. At least, those were in fact the differences in my programs at my alma mater.
even then, introductory assembly was a 200 level course at my school. we did in fact learn how assembly code worked to help better prep us for advanced software courses where we manage memory and develop domain knowledge.
More seriously everyone knows there are patterns that keep repeating themselves from low level on up. The more exposure you get the better you get at understanding them. I think there's a reasonable question about if the best methodology for understanding is to see the same thing repeatedly (that is just study JavaScript) or to see multiple variations on the same thing (that is learn the entire stack imperfectly).
The counter to this in my mind is that learning for it's own sake is often imperfect. If I know I'm learning about semiconductors simply to get on to learning Assembly and then JavaScript (my actual goal) I'll retain less then if the initial steps are the goal itself.
Pedagogy is a tricky thing.
Using AI allows me to take and feed my notes to it and have it create multiple choice quizzes. Same for assignments where my teacher asks us to define terms. So far doing this I am able to retain things quickly and speak intelligently during the next class.
So dont read the textbook, just learn what is needed and graduate. Mandami is definitely right on this as my new way of learning will get me A(s) and B(s) but im thinking i will miss a lot of important pieces of information.
Intern quality at my FAANG has been dreadful for the past year, and new grads aren't much better. Most don't pass probation nowadays.
Not all math classes yes, if your are learning basic arithmetic yes calculators should not be allowed, but if am learning say applied statistics I would be wasting time adding numbers by hand and not learning what I am actually supposed to be taking the course for. A tool is just a tool[1].
Even when it was not allowed, schools would still teach you how to use a calculator correctly and your peers that 80085 looks like BOOBS. Effective AI usage education specifically as an aid to learning is important for children today. There is no avoiding the fact that generative models will part of their whole lives .
The instructor or the school board with inputs from professional experts in learning pedagogy should be taking this kind of call at individual course level. IMO it shouldn't be the mayor doing a blanket ban
[1] There were time when instead of calculators, you would need to carry Clark's Tables booklets you weren't expected to learn numerical analysis to just use value of ln(n) , Clark's tables were a tool, so was the calculator, or a computer or now a machine learning model.
If there's learning how to think stuff going on I'm not sure I want the schools doing that, it seems more like the job of parents.
If I had ChatGPT as a kid I would have learned nothing.
Higher homework scores, but _substantially_ worse test results.
About what you'd expect, right? Friction is important for learning. Make the process too smooth and it slides right off the brain.
a great harm in human society is overconfident extrapolation, especially when treating humans as analogous to weather elements
however, there is literature that explores this aspect, but not relative to what we know right now (and if it is a global trait, does it matter at the general population level?)
Maybe if there are education-tailored AI models strictly limited to explaining and never giving the answer, but even then, how will you know when it's wrong?
I don't know what level of digital tools is best for which types of kids, but the public school system would be prudent in adopting the learnings from the private school system and making sure the best education isn't just available to the rich. The current mayor is mad-effective at his job but does have a habit of doing the ol' wink-wink nod-nod to keep his base happy. This (along with a pointless Waymo ban) sounds like one of those actions.
So on the off-chance one of these tools (that even used AI in the background) could create a tool that actually worked well for, say, gifted public schools kids that don't need as much one-on-one attention, they could free up limited human teacher hours to give weaker kids more attention, that is no longer a possibility.
That possibility, only remains for private school kids, who don't even need it in the first place.
"New York City Bans A.I. in Elementary and Middle Schools"
First sentence is "A new policy announced by the Department of Education bars younger students from using artificial intelligence, but stops short of a moratorium."
Not sure how handfuloflight got "Mamdani Bans AI in NYC Schools" from that. I'm also seeing zero instances of "Mamdani" in the article?
And several million more immigrants than the entire country of Switzerland.
We know what works in education. It's phonics and direct insturction, basically. But nobody gets rich that way so you end up buying into a worse system through two steps:
1. Education PhDs need to justify their existence. They do by reineventing what works. This is basically how we came up with sight words to replace phonics, which is largely responsible for destroying literacy. Whoever comes up with these schemes gets to get academic cachet for being a thought leader. They get invited to all the cool conferences and parties; and
2. Rent-seeking edtech companies end up creating solutions for things that aren't really problems. This long predated AI. It suffers from all the normal issues of corruption in government procurement. Politicians and superintendents are enticed ("bribed") to buy solutions that teachers are then mandated to use, forcing students onto devices to work a program with dubious results.
You see this pattern play out in a ton of government R&D. For example, SMR reactors in nuclear power. There's no world that SMR makes sense. Nuclear plants benefit from scale. But behind all this you'll find a bunch of physics academics who pushed this idea to get status, some of whom then get hired by SMR startups and others get involved in government procurement. Everybody is grfiting.
So while I welcome schools not using AI early on, that's barelly scratching the surface. And that begins with teaching literacy and holding students back until they can read sufficiently for their grade level.
> But nobody gets rich that way so you end up buying into a worse system
As an example Two Minute Papers used to to be great but these days he's mostly covering AI advances, which to be honest become boring really quickly.
IMO the _material_ being taught is secondary; what truly matters is teaching kids to think, to be bored, to engage with something they don't love but do a good job anyway, to find things they _do_ love that might exist on the other side of some amount of difficulty. They need to be taught how to think, and to have some grit.
https://www.yahoo.com/news/us/articles/silicon-valley-parent...
https://www.nytimes.com/2026/08/18/technology/silicon-valley...
I think we already know the answer to this question, I’m not sure what America plans to do about it, or maybe just fade away into irrelevance?
AI feeding students the right answers is damaging, but feeding them the right questions at the right moment is the way forward.
It takes years to understand the effects of the changes we implement in our school systems and curriculums and this attitude of change it now and hope for the best later is what's gotten us to a point where high school students can't read at a high school level (and this problem was happening before LLMs).
Your hypothetical utopia of education should not be tested out on the masses.
At some point they will begin running out of human created data for model training and schools with strict AI policies will be the best source of non AI data going forward.
Thats kind of a symbiotic relationship that makes sense for students and AI companies.
Once AI output begins replacing human output and the models begin training on it, there may be increasingly few places left for data that isnt likely to lead to a model collapse.
Agency is key.
Kids can still access and use AI outside of school meaning the enforcement is entirely on the school and doesn't require massively a privacy invasion for adults by way of age verification. Schools already put children under (from an adult perspective at least) a pretty intense rules regime so this doesn't represent any kind of change—indeed it's probably better framed as a lack of a change from the status quo.
Now, 20 years later, I see the underpinnings of this being formed in real time. I did not heed the warnings.
I never cheated, but it's far too high a bar to expect adolescents to be able to restrain themselves from a magic answer machine.
And the more they use it, the further they're going to fall behind in the basics (being able to do math, read, and write).
https://dn710709.ca.archive.org/0/items/TheFeelingOfPower/Th...
> 75.3 percent were economically disadvantaged [0]
75.3% = "if their family participates in specific economic assistance programs or meets distinct low-income thresholds."
[0] https://www.schools.nyc.gov/about-us/reports/nycps-data-at-a...
Yes, that's my issue. The study is focused on grades and I'm more concerned about deeper logical thinking. You don't get the latter without some friction, and that let's you build up for skills needed in high school and beyond.
The mechanical ability of typing 1+1 into a calculator isn't my concern. It's the ability to know why you're typing 1+1 in as an intermediate step in solving real problems.
Even allowing for generous overhead a ridiculous small class size of 15 could afford $320 000/year salary.
Its not the money.
Not having the skills to assess the accuracy / validity of the information you accessed at breakneck speed however is going to be the downfall of us all.
Ai just made it easier to generate misinformation faster. We were already suffering from the blight of misinformation spreading through various media channels.
If AI can be used to better teach effective media and information literacy, we should be rolling that out in schools.
But it sounds like AI in public schools is largely just being used for speed running homework.
I think people conflate productivity for learning. LLMs can greatly help with productivity. But... so can hiring an employee. Well, what have you learned? Nothing, you outsourced 90% of it. You decidedly learned less than if you were less productive and did it yourself.
Someone gave it to me about 2/3's of the way through Calc 1. I truly regret having become dependent on it. I studied enough to know roughly what I was doing but not enough to not be dependent.
I suspect learning with AI is the same sort of thing. You're learning, but you're not deep learning .
AI has been immensely valuable, but I suspect only so because I am intrinsically motivated to learn this material. When I make errors in solving problems due to a confusion in understanding, it can help me work through the problem the same way a TA would, without having to wait and where I can ask a million clarifying questions until the idea is concrete in my mind.
Could I abuse it to just answer questions for me? Absolutely. But since that’s at odds with my actual goal, I don’t.
I don’t know how to solve an integral, but I understand that an electricity meter or flowmeter (gas and air) is performing continuous integration to keep track of total power usage or fluid flow. That’s good enough for me, and if I had to I could approximate a solution with a piece of paper, a pair of scissors, and a reasonably accurate scale; or trapezoids and graph paper.
Re: your edit, calculus would definitely come in handy if you had an interest in RF. The good thing is, you could still learn it :)
If you don't know math, you often won't be able to even realize when math is needed.
There's a Steve Yegge quote about this:
> You'll gradually realize that your math deficiencies aren't just something that you might need to beef up on if you ever "need to"; you'll see that virtually every problem space has a mathematical modeling component that you were blissfully unaware of until Done, and Gets Things Smart gal points it out to you and says, "There's an infinitely smarter approach, which by the way I implemented over the weekend." You stare slack-jawed for a little while, and then she says: "Here's a ball. Why don't you go bounce it?"
You're right that most people don't need to know calculus in their lives, but to my knowledge calculus also isn't generally required to get through college/university for most people, so it's fair to assume that the people in the course are there because they need or want to learn it. When I was a TA for calculus, most of my students were pre-med, engineers, or in other sciences.
I've done work on fluid simulations. It sure would have been nice to understand the buttons being pushed so I can better validate what's going on.
And I think I did well in calculus. Just not "mastery of multi-variable calculus to transform fluid dynamics formulas into software" well.
There's also the issue that modern teachers are more like exam coaches rather than trying to actually teach kids.
AI is just the weird guy that sneaked its way in, without any reference. That is know to lie often in a very convincing way, specially when it does not have an answer, and that can be persuaded to include a lot of weird random ideas or fake facts in their classes.
Okay. Let’s revisit this in five years. In the meantime, let’s not experiment on our kids.
There's good teachers and bad teachers, there's good AI implementations and bad ones. This bans all of them.
I'm not particularly worried, meaningful experimentation will continue in private schools. I just hope public schools can learn from that and open it back up when it's mature.
This was the same argument made for smartphones in schools.
> it cannot wait until after they graduate high school
I have a strong opinion about technology in early education. The evidence is more mixed for high-school students—I don’t have a view there.
If the argument is no AI in elementary or middle school, and controller introduction in high school, I could get on board with that as long as the solution is being developed by the state and not one of the AI companies or tech giants.
> everyone needs to learn to work with AI to do advanced math
Most people don’t need to know advanced math. Our problem, currently, is college students being unable to do basic math. Or hold attention through a book.
> it has to be both
Why? AI isn’t a mature technology. I’m not convinced someone who was chatting with ChatGPT in 2023 is more productive with AI, today, than someone who started building agentic workflows into their processes six months ago.
I'm curious, how do you envision this development happening? You're (presumably) calling for all school boards to ban any AI developed anything in classrooms like NY just did.
Are you expecting public universities to perhaps carry out studies first?
For reference, I think putting a tablet with an AI in front of kids to talk to is a silly boogeyman most people seem to be talking about in this thread. My vision was always teacher tooling that would greatly expand the teachers ability to develop individualized curriculum and test a kids progress in a classroom etc. But calling for such blanket bans means we can't even test any of this..anywhere.
Idk. I don’t have a great solution. But I also don’t think kids being slightly less familiar with AI at schools is a real problem.
> how do you envision this development happening?
Literally two classrooms I’ve seen in the Bronx in the last 6 months. Tablets are cheaper than teachers.
> My vision was always teacher tooling
I like this. I support it. I don’t support putting it in front of kids. Given how corrupt our AI leadership is right now, erring on the side of caution seems prudent.
> calling for such blanket bans means we can't even test any of this
No? We still have AI everywhere. We can still do controlled studies.
So... talking about future capabilities.
>Kids are going to need to engage with this directly
Just to get this straight: are you the kind of person to give your 1st grader a calculator and say "just use this. you won't ever have to add numbers by hand anyway"? So much metaphors of AI go around calculators, but this is a problems we solved decades ago.
>everyone needs to learn to work with AI to do advanced math
We have so much complaints over the application of math as is. Now you're arguing that it's useful again because we can offload it to an AI?
I think banning smartphone/computer use before 3rd grade in the classroom sounds like a good idea.
Math is incredibly useful. I think the biggest problem with math is that people don't understand how to apply it, it's not that applications are hard to find. AI actually explains how to apply it very directly. I can never remember how to calculate compound interest, an AI can do it instantly.
Mamdani's banning teachers from using AI to grade assignments.
> So... talking about future capabilities.
No, I'm talking about current capabilities. The studies that prove that current AI is better at grading assignments than humans will take at least a year, if not several.
The question is if we change what assignments teachers give, or if we accept that grading is a task best done by computers. If we change the assignments that teachers give, I don't know what that looks like. Regardless, there's going to be a lot of experimentation going on. Banning AI doesn't help.
I started secondary school (British closest equivalent to highschool) in autumn 1995. One of the teachers opined that they liked Apple's computers "but no point teaching them because they'd be bankrupt soon". Took a while for the school to replace the RISC OS machines with Windows.
By 2000, when I started my A-levels (UK has two years between secondary school and university for this qualification), Windows and MacOS had changed so much as to make any lessons from the start of secondary school, on RISC OS, irrelevant.
Five years later, 2005/06, I noticed while using my laptop to record my final year of university lectures (I was on a four year course), that most of us in the lecture hall had gone from pen and paper in the first year to laptops in the final year.
Five more, 2010, smartphones had no mice, and the on-screen keyboards were significantly different from physical keyboards. Multitasking was a very different experience compared to a desktop or laptop OS. Social media was replacing "never tell anyone on the internet your real name" with "tell everyone on the internet even the most banal details about your life".
Five more, 2015, and smartphones and apps were shifting from exceptions to expectations.
Five more, 2020, we're walking into rooms and telling the always-listening microphone to adjust the lights or play some music, and using the now-free video calls we'd come to expect as a minimum requirement of any chat service to get us through the isolation of the pandemic.
Two years later, 2022, people started cheating on their homework with an AI which knew more than they did, even despite how bad early AI was at maths in particular.
Four years later, 2026, the AI isn't just helping kids cheat, it is solving (a limited subset of) previously unsolved maths problems at research level.
Even last year, 2025, when it was "merely" doing well on Olympiad-type challenges, I didn't even understand the questions. My A-levels included maths and further maths, which isn't a boast because I didn't follow this up with a maths degree, but it should give a sense of how much further ahead of mandatory schooling AI already was even last year.
None of the headline grabbing CEO's have any official education relationships. This seems like a "I saw some bad headlines" vibey reaction.
>>No? We still have AI everywhere. We can still do controlled studies.
Not in schools we can't, where Teacher productivity has slumped. Teachers are spending more time doing admin work, and less time with kids. Meanwhile, rich parents who send their kids to Private schools can experiment ASAP (this ban doesn't affect them), and get any and all productivity enhancements. Shouldn't the government be creating some space for experimentation? Shouldn't it be the onus of the banning party to create this space?