Table of Contents
The School AI Ban Debate: What Chicago's Pledge Would Do
Half of Chicago's school board candidates backed a three-year ban. What the school AI ban debate would mean for students, and why a pledge is not a policy.
The school AI ban debate stopped being hypothetical in Chicago on September 10, 2026, when roughly half of the candidates running for the Chicago Board of Education signed a pledge backing a three-year ban on AI use by Chicago Public Schools students and educators. The same pledge called for no screen time before third grade and an end to biometric data collection in schools. A pledge is a campaign document, not a board policy, and that gap is the single most useful thing for a CPS parent to understand right now. But the three planks tell you exactly where this is heading, and two of them would reach far more families than the AI plank.
Key Takeaways
- A candidate pledge binds nobody. CPS policy changes require a vote by the seated Board of Education, and a signature in September 2026 is a statement of intent.
- The pledge has three parts, and the no-screens-before-third-grade plank would affect more CPS families day to day than the AI ban.
- An AI ban in practice means network-level blocks on district devices and accounts, plus staff rules. It does not reach a child’s phone at home.
- Two of the three largest U.S. districts already restrict student-facing AI: New York City for Pre-K–8 as of September 2, 2026, and Los Angeles Unified on district devices.
- The strongest evidence does not say AI helps or hurts. It says unguarded AI hurts and well-designed AI helps, which argues for configuration rules rather than a switch.
What the September 10 pledge actually says
A pledge in a school board race is a public commitment by candidates to vote a certain way if elected. It creates political accountability, not legal obligation.
According to the AI-in-education policy tracker entry dated September 10, 2026, roughly half the candidates in the Chicago Board of Education race signed a pledge calling for a three-year ban on AI use by CPS students and educators, no screen time before third grade, and an end to biometric data collection.
Read that list again, because it is three separate fights wearing one coat. The AI ban is the headline. The pre-third-grade screen limit is a curriculum and staffing decision that touches every K–2 classroom. Biometrics covers fingerprint scanners in lunch lines, face recognition on building cameras, and voice data. Those three planks would be implemented by different departments on different timelines, and a board could easily adopt one and drop the others.
What would have to happen for any of it to bind your child: the candidates win, take their seats, draft a policy, put it on an agenda, and vote. Then CPS operations has to implement it, which for an AI ban means the IT department changing network filtering and account permissions. For a screen-time limit it means re-equipping primary classrooms and retraining teachers. The first step is cheap. The second is not.
A ban, a block, and a rulebook are three different things
Districts have converged on three distinct instruments this autumn, and they are frequently confused in conversation because headlines treat them as one.
A moratorium is time-limited and student-facing. New York City Public Schools adopted one on September 2, 2026: a one-year ban on student-facing generative AI for Pre-K through grade 8, affecting roughly 600,000 students, plus a district-wide bar on companion chatbots at any age, screen-time limits, and twice-yearly AI literacy classes. High schoolers are outside the ban but reach AI only through supervised pilots.
A device block is technical and quiet. Los Angeles Unified disclosed on September 2, 2026 that software already blocked all student-facing generative AI on district-issued devices, reversing an earlier policy that had allowed students 13 and older to use AI after digital-citizenship instruction. Note the sequence: the block was in place before it was announced.
A rulebook sorts uses rather than banning them. Washington DC’s Office of the State Superintendent of Education released a nonmandatory stoplight framework on September 1, 2026 that labels uses approved, caution or prohibited. It bars AI from grading, discipline, IEP decisions and teacher evaluations while permitting lesson planning and messaging, with final judgment left to the teacher.
| Instrument | Jurisdiction | Who it covers | What it actually does | Binding now? |
|---|---|---|---|---|
| Chicago pledge (Sept 10, 2026) | Chicago Public Schools | Students and educators, if adopted | Proposed 3-year AI ban, no screens pre-grade 3, no biometrics | No. Campaign pledge |
| NYC moratorium (Sept 2, 2026) | NYC Public Schools | ~600,000 Pre-K–8 students | One-year ban on student-facing generative AI; companion chatbots barred district-wide | Yes |
| LAUSD device block (Sept 2, 2026) | Los Angeles Unified | Students on district devices | Software block on student-facing generative AI | Yes |
| DC stoplight framework (Sept 1, 2026) | DC schools | Staff practice | Sorts uses; bars AI from grading, discipline, IEPs, evaluations | No. Nonmandatory |
The pattern worth noticing: the districts that acted fastest acted on the youngest students and on companion chatbots, not on high school research tools. That is a defensible reading of the evidence, and it is a different policy than “ban AI.”
What the research supports, and what it does not
Here is the honest state of the evidence, which does not hand either side of the school AI ban debate a clean win.
Bastani and colleagues published a large field experiment in PNAS in 2025 using GPT-4 with high school math students. Students did better during practice sessions with AI help, then showed reduced skill development once the AI was taken away. The authors also found the effect was mitigated by design choices, specifically configuring the tutor to deliver teacher-designed hints instead of answers. That is a finding about guardrails, not about AI as a category.
Kestin and colleagues reported the other direction in Scientific Reports in 2025: physics students using a purpose-built AI tutor that followed the same pedagogical principles as the classroom version “learn significantly more in less time” than peers in in-class active learning.
On the cognitive side, Kosmyna and colleagues at MIT Media Lab measured EEG activity during essay writing across three conditions and reported that participants using a language model showed the weakest brain network connectivity, reported lower ownership of their essays, and struggled to recall their own sentences. It is a preprint with a modest sample, and it measures a single task. Treat it as a signal, not a verdict. Separately, a 2025 paper in Societies by Michael Gerlich of SBS Swiss Business School surveyed 666 UK participants and found a strong negative correlation between AI tool use and critical thinking scores (r = −0.68), mediated by cognitive offloading. Correlational, self-reported, adults.
Put together: a blanket ban is a blunt instrument aimed at a real problem. A configuration rule, the DC approach, matches the evidence better. Neither position is unreasonable, and anyone claiming the research settles it has not read it.
The plank almost nobody is discussing
No screen time before third grade would change more CPS classrooms than the AI ban, and it is the plank with the least coverage.
Most primary classrooms now run some portion of literacy and math practice on devices. Remove that and you need paper materials, more adult hands for small-group work, and a new assessment plan, because much of the data teachers use to group K–2 readers comes from the software. This is not an argument against the plank. Several districts have moved back toward print in early grades, and the National Assessment of Educational Progress reports that the share of 13-year-olds reading for fun almost every day fell to 14% in 2025 from 27% in 2012, which is the kind of number that motivates exactly this proposal.
It is an argument for asking a specific question: what replaces the software, and who pays for it? A screen ban without a materials budget becomes a worksheet ban in practice, because the worksheets were never ordered.
What to do if you are a CPS parent
Separate the three planks when you ask
Email your board member or candidate and ask about each one on its own: the AI ban, the K–2 screen limit, the biometrics plank. Candidates who signed a bundled pledge may hold very different views on the parts. The answers will tell you which plank has real momentum.
Find out what is already blocked
Most districts already filter more than parents realise, as LAUSD’s September disclosure showed. Ask CPS IT, through the school office, which categories of sites and tools are blocked on student accounts today. You may discover the ban you are debating is substantially already in effect, which changes what the vote would accomplish.
Ask the biometrics question in writing
Biometric data is the plank with the clearest paper trail. Ask whether the school uses fingerprint or palm scanning for meals or attendance, whether building cameras run any recognition software, and whether any vendor stores a biometric template. Request the vendor names. This is a records question, and records questions get answers.
Plan for the gap between policy and practice
If a ban passes, your child’s school AI exposure moves from district devices to personal devices, where the district has no filter and you have all of it. Common Sense Media’s 2025 research found nearly three in four U.S. teens had used an AI companion, half of them regularly. A district ban does not touch that number. Our comparison of district AI policy models covers how ban, scaffold and open approaches play out at home.
What not to do
Do not treat the pledge as settled policy in either direction. Parents who assume the ban is coming stop asking about tool configuration, and parents who assume it will fail stop showing up to meetings. The useful stance is to track the agenda. A policy appears on a board agenda before it is voted on, and that agenda is public.
What to Watch For Over the Next 3 Months
- Week 4: Check whether any pledge language appears in a drafted board policy. Pledge-to-draft is the conversion that matters, and it usually happens within the first two months of a new board’s term.
- Month 2 red flags: A ban announced without an implementation date. A screen-time plank adopted without a materials line in the budget. Any claim that CPS “has no AI in schools,” which is unlikely to be true of teacher-side tools.
- Month 3 self-check: Can you state, in one sentence each, what the district currently blocks, what it currently permits, and what it currently collects about your child biometrically? Three sentences. If you cannot write them, that is the homework.
Frequently Asked Questions
Does the pledge mean AI is banned in Chicago schools?
No. It is a campaign commitment signed by roughly half the candidates on September 10, 2026. A ban would require the seated Board of Education to adopt a policy by vote, and then CPS operations to implement it.
Would a ban cover my child’s phone?
No. District policies reach district devices, district accounts and district networks. A personal phone on cellular data is outside that perimeter entirely, which is why the home conversation matters more than the policy in most families.
Which big districts have actually restricted AI?
New York City adopted a one-year ban on student-facing generative AI for Pre-K–8 on September 2, 2026, covering about 600,000 students, and bars companion chatbots district-wide. Los Angeles Unified blocks student-facing generative AI on district devices. Our breakdown of the NYC ban covers the detail.
Is there evidence that banning AI improves learning?
Not directly. The evidence addresses AI use, not bans. Bastani and colleagues found unguarded GPT-4 access reduced later skill, while Kestin and colleagues found a purpose-built tutor outperformed active learning. That pattern supports rules about how tools are configured more than it supports a switch.
What does “no biometric data collection” cover?
Typically fingerprint or palm scanning for lunch and attendance, face recognition on building cameras, and voice data. Ask which systems your school uses and which vendors store templates, because the answer is usually specific and documentable.
My child is in second grade. Which plank should I watch?
The screen-time plank. A limit on device use before third grade would change daily instruction in your child’s classroom more than any AI rule, and the question to ask is what print materials and staffing would replace the software.
About the author
Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.
Sources
- Pursuit. (2026). “AI in Education: News, Policies, Innovations” (entries dated September 1, 2, and 10, 2026). https://www.pursuit.us/news/ai-in-education-news-policies-innovations
- Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakcı, Ö., & Mariman, R. (2025). “Generative AI without guardrails can harm learning: Evidence from high school mathematics.” Proceedings of the National Academy of Sciences. https://doi.org/10.1073/pnas.2422633122
- Kestin, G., Miller, K., Klales, A., Milbourne, T., & Ponti, G. (2025). “AI tutoring outperforms in-class active learning: an RCT introducing a novel research-based design in an authentic educational setting.” Scientific Reports. https://www.nature.com/articles/s41598-025-97652-6
- Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X.-H., Beresnitzky, A. V., Braunstein, I., & Maes, P. (2025). “Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task.” arXiv:2506.08872. https://arxiv.org/abs/2506.08872
- Gerlich, M. (2025). “AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking.” Societies, 15(1), 6. https://www.mdpi.com/2075-4698/15/1/6
- National Center for Education Statistics. (2025). NAEP Long-Term Trend Assessment results. The Nation’s Report Card. https://www.nationsreportcard.gov/
- Common Sense Media. (2025). “Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions.” https://www.commonsensemedia.org/research/talk-trust-and-trade-offs-how-and-why-teens-use-ai-companions
