🛝 Changing the Rule Was the Easy Part

Welcome to Playground Post, a bi-weekly newsletter that keeps education innovators ahead of what's next.

This week's reality check: New York City says it is disabling AI features in software covered by 38 citywide contracts, without a complete picture of classroom use. Michigan has a curriculum-notification rule but no required curriculum-reporting mechanism. Meanwhile, California's later school starts offer encouraging sleep evidence.

Bring one big idea. Leave ready to move it forward.

Have an idea you keep thinking about - but haven’t had the time, space, or framework to really explore?

Spend one high-energy day at a Design Dreaming Workshop, where you’ll dream bigger, pressure-test your thinking, and turn your idea into something you can actually move forward.

Leave with greater clarity, 2–3 concrete next steps, and a practical innovation framework you can use again and again.

NYC, October 16 | D.C., October 20 | Chicago, October 22

Playground Post readers get 25% off any ticket.
Use discount code: D5OB8A

Early-bird tickets are $395 through September 18.

Learn more + reserve your spot →

💎 Data Gem

Only 14% of California's estimated 86,000 children ages 0–3 experiencing homelessness were enrolled in early-learning programs in 2023–24. These are estimates from forthcoming SchoolHouse Connection and University of Michigan research in the report on early-learning access.

The AI Ban Came Before the Inventory

New York City has told schools to stop student-facing generative AI from preschool through eighth grade.

Now comes the harder question: where is it?

Officials acknowledged an incomplete picture of use across 1,600 schools, according to the explanation of the new policy.

The city said AI features would be disabled in software covered by 38 citywide contracts. Vendors sought clarification or disputed whether their products contained the affected features.

A product name is not enough to settle that question. The policy distinguishes who uses a feature and what they use it for.

Teachers can use AI for planning and translation, but not for grading, behavior monitoring, or developing IEPs. Exceptions remain for specified disability and English-learner supports and required assessments.

Purchases below $25,000 now face additional review.

At the high-school level, five pilots may reach up to 50,000 students.

This is a change in operating rules, not evidence that a ban has improved learning. Enforcement and product-level details remained unresolved when the policy was reported.

The immediate task for a technology team is therefore unusually specific. It must distinguish an allowed teacher workflow from a prohibited student feature, sometimes inside the same product.

That calls for something more useful than a yes-or-no label beside a vendor's name. A review should be able to identify the feature, its users, the permission setting, and any documented exception.

For education innovators, the opening is feature-level software inventories, grade-and-role access controls, and procurement review services that preserve evidence of the configuration a school approved. Let an IT administrator verify restrictions without accidentally removing an allowed accommodation. Make a policy decision testable in the software students actually open.

The Curriculum Rule Had No Reporting System

Michigan can withdraw state funding from a district that uses an unvetted literacy curriculum and fails to tell parents.

But the education department has no mechanism requiring districts to report which curriculum they use.

State law limits its ability to require new reports without authorization, according to the reporting on Michigan's literacy reforms.

The notification rule began in 2025–26.

No district had lost funding for noncompliance when the article was published.

That does not prove every district complied. It identifies a limit in how the state can check.

The new achievement results explain the urgency. Third-grade English proficiency fell to 38.2%, from 38.9%, in spring 2026.

Still, the results cannot be treated as a verdict on reforms that have not fully arrived. Evidence-based reading instruction and dyslexia-screening requirements take full effect in 2027–28. The elementary-teacher training deadline is 2031–32.

The distinction is important for educators evaluating the next promise of a better curriculum: selection, notification, training, and classroom use are separate steps.

A list of approved materials answers which products passed a review. It does not, by itself, answer which edition a school bought, what teachers received, or whether families got the required notice.

Those are useful questions for an implementation review, not findings the test scores can supply.

For education innovators, the practical opening is curriculum inventory software, parent-notification workflows, and implementation audits that connect adopted materials to training and classroom delivery. Make it possible to distinguish a missing record from a missed requirement, with an accountable person for each follow-up. Give schools a usable trail from district choice to teacher support.

The First Bell Was Part of the Problem

A school can invest in lessons and still make learning harder with its timetable.

California offers a useful test of that possibility.

An NBER working paper examined the state's later-start mandate, which generally requires middle schools to begin no earlier than 8 a.m. and high schools no earlier than 8:30 a.m.

The researchers found a 13% increase in the share of students sleeping at least eight hours, according to the report on the findings.

That is a relative increase, not 13 percentage points. The research also found eighth-grade gains in math and English.

Researchers used teachers' commute arrival times as a proxy for school start times, an important limitation when interpreting the result.

Then there is the implementation problem: buses, family work schedules, and after-school activities have to fit the new day.

The reporting gives a separate operational example.

San Francisco redesigned bus schedules across more than 120 schools, saving about $5 million with help from university operations experts.

Those savings are not an outcome estimate from the statewide sleep study, nor a promise another district will save the same amount.

They do suggest a more productive planning question than whether a later bell is inconvenient. Which combinations of routes and start times protect the intended benefit while limiting the disruption families have to absorb?

For education innovators, the opportunity is bell-time and bus-route scenario planning, family schedule consultation, and implementation services that test options together before a change takes effect. Show the tradeoffs plainly: pickup times, dismissal times, staffing needs, and who faces a difficult transition. Help a district deliver the additional sleep its policy is meant to create.

⚡️More Quick Hits

This week in education:

• 77% of teachers feel optimistic, even as nearly one-third report high or critical stress — A Teachers Pay Teachers survey of nearly 7,500 educators found that a hopeful start to the year can coexist with substantial pressure.

• 37 federal school-data files arrive without the usual public-facing summary — The 2023–24 Civil Rights Data Collection adds information on teacher retention and preschool screening, but researchers were still processing the release and its accessible First Look Report was absent.

• Two Indiana districts plan to end bus service if tax measures fail — Tri-Creek, with roughly 3,300 students, and Mississinewa face transportation cuts after state revenue changes, making November's funding decisions consequential for school access.

• US 15-year-olds lost 14 reading points on the latest PISA — American math and science results were statistically unchanged from 2022, while the reading decline adds urgency to support for struggling adolescent readers.

To stay up-to-date on all things education innovation, visit us at playgroundpost.com.