Orange County Public Schools asked teachers and parents what to do about artificial intelligence, and roughly 20,000 of them answered. On August 11, the first day of classes in central Florida, the district’s response took effect.
The policy runs a few pages. Its most interesting line concerns what teachers are not allowed to lean on: AI-detection software. Orange County does not require it, and a teacher who suspects a student handed in a chatbot’s work has to make that case some other way. After three years of districts buying detection tools and then quietly discovering how often they flag the wrong kid, that is a notable place to land.
Florida sets the tone
The Orange County school board approved the policy on July 28, two weeks before students returned. The long consultation is the part worth paying attention to. Twenty thousand survey responses is not a rubber stamp, and the resulting document reads less like a ban than a set of ground rules for a tool the district has accepted is already in the building.
Charleston County School District in South Carolina reached a similar conclusion this month, rolling out its first districtwide AI policy with the stated goal of teaching students and staff to use the technology responsibly rather than keeping it out of classrooms. Two districts, two states, the same basic calculation.
The number that ended the debate
Common Sense Media’s 2026 census found that 86 percent of American children and teenagers aged 9 to 17 had used or interacted with generative AI. Among those who use it, 85 percent said they use it for schoolwork.
Those figures explain the shift better than any policy paper. A district writing a prohibition in 2026 is writing a prohibition that almost nine in ten of its students are already ignoring. Enforcement was never going to scale. What districts can still control is whether students learn to check the output, cite it honestly, and recognise when the machine is confidently wrong.
Boston tries something harder
Boston Public Schools has gone further than a usage policy. Starting in September, every high school in the district begins an AI literacy programme built over the summer by teachers and industry partners, with the goal that students graduate fluent in the tools rather than merely permitted to use them.
The structure is worth describing. Teachers from each of the district’s roughly 25 high schools were selected as AI ambassadors and trained over the summer. Students will also be able to take credit-bearing AI courses at UMass Boston. The programme was seeded by a $1 million donation from the technology entrepreneur Paul English, who co-founded Kayak.
Boston has been widely described as the first major-city district in the country to commit to AI fluency for every high school graduate. Whether it hardens into a formal graduation requirement remains open; the district has framed it so far as a proficiency goal. Either way it is a bigger swing than a set of rules, and it costs considerably more.
Statehouses are catching up
FutureEd counted 77 bills across 27 states in the 2026 legislative session dealing with artificial intelligence in classroom instruction. Four more states now require districts to adopt an AI policy at all, a low bar that a surprising number of districts had not cleared.
Idaho’s S.B. 1227 is a reasonable template for what these laws look like. It directs the state education department to build a generative AI framework covering privacy, procurement safeguards, transparency, academic integrity, AI literacy standards and professional development, and requires districts to adopt aligned policies. The interesting word there is procurement. Districts are not only deciding what students may do with AI, they are deciding which vendors get access to student data, and most of them have never had to evaluate a model provider before.
What to watch this term
The policies are written. The harder part starts now, and it lands on teachers who got a summer workshop and a slide deck. Volusia County built an AI academy to run professional development on district-approved tools for the 2026-27 year. Neither that nor Orange County’s training can guarantee consistency across every classroom, and consistency is exactly what an academic integrity rule needs to survive a parent appeal.
The first real test arrives around the end of the first grading period, when someone contests a grade over suspected AI use and the district has to show its work without a detection score to point at. How that goes will shape the next round of policy far more than anything approved in July.
For more coverage of AI in education, visit Mylistingo.
The Netherlands, for internationals
One email a week: the week's news for internationals in the Netherlands, new practical guides and what changed in the cities. No spam, unsubscribe any time.
