AI in schools: Charleston County wrote its own policy for 50,000 students
Charleston County School District wrote an AI policy for roughly 50,000 students on its own initiative, while the AP reports that 37 states have published AI guidance for schools that districts can build from.
Key Takeways
- Charleston County School District wrote an AI policy for roughly 50,000 students on its own initiative, with teachers, students, and parents helping draft it.
- The AP reports that 37 states have published official AI guidance for schools that districts can use as a blueprint.
- Utah created a full-time state AI education specialist role in 2024, has trained more than 7,000 teachers, and requires every district to have an AI policy in place by July 2027.
- Curriculum half-life, meaning how fast a syllabus goes stale once the tools underneath it change, is what these state and district decisions are managing.
Charleston County School District built its own AI policy for roughly 50,000 students on its own initiative, with teachers, students, and parents helping write it. Student focus groups shaped the language. A second phase, training staff on how to use the policy, is rolling out now. The AP reports that 37 states have published official AI guidance for schools that districts can use as a blueprint. South Carolina has yet to publish one, so Charleston County did the work locally. That is what AI in schools looks like once the argument about whether to allow it has been settled.
Jocelyn Gecker reported this for the AP on August 24, and the pattern she describes is a national one. American K-12 districts are moving away from blanket bans and toward structured AI literacy instruction. Code.org and Gallup research cited in that same reporting found a majority of teenagers and a majority of teachers already using AI for schoolwork. Deputy superintendent Lucas Clamp is among the district officials in the piece describing how the policy work actually gets done.
Utah turned AI literacy into a job description
Utah created a full-time AI education specialist position inside state government in 2024, and Matt Winters holds it. More than 7,000 Utah teachers have been trained since. The Utah State Board of Education now requires every district in the state to have an AI policy in place by July 2027. That hands districts a date to plan against instead of a mood to react to. Kristina Yamada, a Digital Technology Specialist with the Utah State Board of Education, works on the same effort. Maine, West Virginia, and Georgia have since created the same kind of role.
Funding a permanent state position for this is rare anywhere in the country, and it deserves the credit. A one-time task force produces a document. A staffed, funded role produces 7,000 trained teachers and a deadline other states can copy.
In Utah's Canyons School District, more than 150 AI literacy trainings have run so far. Emma Moss oversees the district's AI work, and a $500,000 grant from a local healthcare provider paid for part of it. A healthcare company paying to train schoolteachers tells you where the pressure is coming from. Employers can already see which skills are expiring, and some of them are willing to fund the repair upstream.
The old approach was a policy about a tool
Watch a 9th grader work through a homework question with a chatbot open in the next tab. Then watch the teacher grading that homework, who has the same tab open for lesson planning. Both of them are working under rules their school wrote for a set of tools that has changed twice since. Common Sense Media put a number on the first half of that scene: 70% of teens already use AI for schoolwork.
Plato recorded Socrates worrying that writing would ruin memory. The tool won, memory changed, and the argument moved to what schools should teach with it.
A ban is a policy about a tool. A curriculum is a bet on what a student will need.
Curriculum half-life is what these states are managing
A syllabus is a wager. A district decides what a 9th grader should know, and that decision has to hold up through four more years of school and whatever waits at the end of them. I call that the four-year bet. Curriculum half-life is how fast the bet goes stale once the tools underneath it change. Charleston County, Utah, Maine, West Virginia, and Georgia are all doing one thing: shortening the distance between when the tools move and when the teaching moves.
This is the through-line across everything I follow on AI in education, and it is happening at every level of the system at once. Florida set an AI policy mandate across 28 public colleges while the Utah State Board of Education was setting a 2027 deadline for K-12 districts. Amanda Bickerstaff, CEO of the nonprofit AI for Education, and Rebecca Winthrop, who directs the Center for Universal Education at Brookings, both appear in the AP piece. The people building curriculum and the people studying it are working the same problem in the same year. Chris Agnew directs a Generative AI for Education hub at Stanford, which is itself a description of how fast this moved.
What this changes for anyone outside a school building
The parent version of this story is plain enough. Your kid's school is deciding, this year, what counts as learning when the tool can write the first draft. The employer version is the same decision with a payroll attached. Every organization has some version of the four-year bet running right now, inside its onboarding, its training budget, and the skills it hires for.
Districts doing this work are doing it with a small number of people and a short calendar. The front end, deciding what AI literacy means for a 9th grader versus a senior, is exactly where outside practitioners can help a deliberative process move faster without taking it over. Charleston County's method, putting students in focus groups before anyone drafted the rules, is one that other districts can lift whole.
Utah's answer took a funded position, 7,000+ teachers, and a date on the calendar. Charleston County's answer took teachers, students, and parents at the same table, plus a second phase for training. Neither answer required waiting for certainty about where the technology finally lands.
So here is the question I am carrying into my own work this month, and you are welcome to borrow it. If you assume that half of what your people learned last year expires within the next two, what is the first thing you would change about how they learn? Ask it at your next planning meeting and watch the response. I am easy to find.
Are students allowed to use AI in school?
It depends on the district. There is no national rule, and state guidance is advisory until a state requires districts to adopt a policy, which Utah now does with a July 2027 deadline. Most district policies set different expectations by grade level and by assignment type, so the same tool can be permitted for brainstorming and restricted on a graded final draft.
Does this shift apply to K-12 only, or to colleges too?
The AP reporting covers K-12. Higher education is moving on its own track, with Florida's mandate across 28 public colleges as one example. The practical difference is that K-12 policy has to account for minors, parental consent, and state student privacy law, so districts usually approve a short list of reviewed tools rather than permitting whatever a student brings from home.
What should a parent ask their child's school about AI?
Ask which tools are approved, what a student is required to disclose when they use one, and who trained the teacher grading the work. The training question is the one most likely to go unanswered, and teacher training is what Utah's 7,000+ number is measuring.
Here is what makes Alex a credible voice on this topic: he advises the California State University system as a member of its AI Working Group, which is the public college version of the AI policy work Charleston County School District and the Utah State Board of Education are doing in K-12 right now.
Bring this question to your team. Book a conversation →
.svg)
