Katy ISD hands out AI access by age: no generative chats until 8th grade

By Zendoric · July 25, 2026.

The Katy Independent School District (Katy ISD), in the Houston area of Texas, will roll out its own framework to regulate the use of artificial intelligence in classrooms in the 2026-27 school year. As Sanée Bell, deputy superintendent of teaching and learning, explained at a school board meeting on July 20, the plan rests on the technology applications TEKS (Texas Essential Knowledge and Skills), the state standards that set how students from kindergarten through eighth grade should interact with technology.

What matters is not that a district "allows" or "prohibits" AI, but that it distributes it by age. In kindergarten through sixth grade (up to ages 11-12), students may only use district-approved platforms, such as Amira Learning or Writable, always under teacher guidance; general-purpose generative chatbots, like Microsoft Copilot or Google Gemini, are off the table. In seventh grade the door opens to approved AI tools, but under supervision and without access to unauthorized generative tools on the school's devices or networks. Only from eighth through twelfth grade—the high school range—can students use generative AI for specific educational purposes, with teacher authorization and an obligation to cite it where appropriate.

The schedule also includes an AI literacy course that will start within a few weeks, as soon as teachers return from vacation, and the publication on the district's website of the full framework and the list of approved tools. Parents will be able to join an AI advisory committee, according to Community Impact, the local outlet that covered the board session.

Broadly, American public schools have been swinging since 2023 between two unhelpful extremes: banning ChatGPT on school devices (a measure that does not prevent its use on a personal phone) or letting each teacher improvise. The "tiered access" model Katy ISD proposes—age as the criterion, closed platforms for the youngest, open and cited generative AI for older students—is a more serious attempt at intermediate governance, aligned with the logic that the ability to use a tool critically is built in stages, not granted or denied wholesale.

Our reading is that the underlying problem is not technological, it is one of execution. A framework like this can make pedagogical sense and still fail for the usual reasons: personal devices beyond the school's control, teachers without enough training to detect misuse, and families with highly unequal resources to complement—or replace—what the school offers. That risk of a two-speed AI literacy, depending on the zip code, is exactly the kind of short-term inequality that should not be downplayed: if AI becomes a basic 21st-century skill, whoever reaches college or their first job without having practiced it in a guided way starts at a real disadvantage.

At the same time, it connects with something we have already argued in our analysis of education as a sector: the teacher who comes out ahead with AI is not the one who bans it nor the one who lets it run free, but the one who orchestrates it, setting out when to lean on it and when not to. That a school district formalizes that gradation—and does not leave it to each classroom's discretion—is a sign that the education system is beginning to treat AI as a competency to be taught, not a threat to be fought. If the approach works and is well documented, it is reasonable to expect other districts in Texas and beyond to replicate it: age-tiered AI literacy stands a better chance of becoming a de facto standard than any outright ban, precisely because it acknowledges what is already obvious to any parent or teacher: the technology is already in children's hands, and the question is not whether they use it, but with what judgment they learn to do so.