Heading into the school year that starts in four days, Mayor Mamdani’s Department of Education and Chancellor Kamar Samuels have announced a generative AI ban for students through 8th grade, limited screen time for younger students, and a slate of approved uses for AI tools for high schoolers that include literacy and career preparation but not the fundamentals of learning. This one-year moratorium isn’t much of a policy, more of a stall.
Samuels came up with this after a series of starts and stops and delays earlier in the year on the use of AI in the schools. But the confusion is not entirely his fault, as AI is spreading rapidly. While clearly we don’t want Anthropic’s Claude writing 5th grade compositions and OpenAI’s ChatGPT solving math homework problem sets, it seems OK to use a spell-checker on an essay that a student wrote herself.
It’s impossible to make a decision of this sort that will please everyone, and indeed it’s inevitable that some will say that this is a betrayal or a stunting of our students’ complete education, but Samuels is at least trying. The evidence that heavy AI use improves the educational experience or education outcomes for our students is very murky and limited. Instead, we have increasingly strong evidence that overuse is stunting their development, which just makes sense; using a calculator for all basic arithmetic will prevent someone from ever learning how to do mental math.
Now apply that to everything else that a student might do in an educational setting, from reading comprehension to reasoning to critical thinking to even making friends. The AI companies themselves have oriented their message and marketing around precisely this, that their systems will do everything and anything for everyone. If that’s so, what’ll be left for our students?
The adoption of this policy came very late in the summer and after a more tentative earlier policy that mainly confused educators and parents. Blame the rapid development of AI tools that has been unbridled and chaotic, with multiple new developments every few months, all carried by a massive wave of hype that is more interested in what AI could one day do than what it actually now does.
This one-school-year moratorium is a good start, though it is insufficient for a DOE that is going to be navigating these treacherous waters for the foreseeable future. We understand the need for a stopgap solution in the interim, but we need permanent and clear guidance for our teachers and administrators as they strive to determine what’s AI buzz and what’s true literacy around tools that will certainly not disappear fully from our lives, but whose role in them remains uncertain.
The proponents of AI tools in all classrooms and at all jobs would have you believe that the mass adoption of these tools is an inevitability, that the only question is of how and how aggressively to adopt language models like ChatGPT and when to introduce them. This is not true; no technology or social evolution is inevitable, and even some young people are already turning away from them.
The main question that ultimately needs to be answered is: what is a K-12 education for? Is it pure preparation to eventually enter a workforce in which AI may well be considered indispensable soon? (Though we should not forget that just a few years ago, coding was considered a crucial skill for learners, and now the entry-level software engineering pathway has collapsed.)
Or is education to teach our students how to think, how to approach problems and figure out solutions and live in society, yes, aided by technology, but ultimately in their own right? We believe it is the latter.