Two AI stories crossed my desk this week. On the surface, they are about the same thing: a school system bringing an AI tool into classrooms. Underneath, they are about something different. They are about how AI adoption in schools actually happens, who gets a say, and what that process does to trust.
In one story, a state education agency quietly connected districts to an AI learning platform that the state’s own board of education had raised serious questions about. In the other, a rural district launched a small, voluntary pilot and told teachers up front that they would be asked what they thought.
Same category of decision. Very different leadership.
Key takeaways
- AI adoption in schools is a leadership decision before it is a technology decision. How a tool arrives shapes whether people trust it.
- A state or vendor recommendation is information, not permission. Districts still have to ask whether a tool is right for their students.
- Voluntary, time-bound pilots with scheduled teacher feedback build credibility. Opaque processes erode it, even when the tool is good.
- Leaders should document how AI decisions are made and share that process with staff, boards and families early.
What’s happening
On October 1, The Texas Tribune and ProPublica published an investigation by Lexi Churchill into how the Texas Education Agency (TEA) and Commissioner Mike Morath helped link public school districts to the AI learning platform behind Alpha School. According to the report, the State Board of Education voted 10-3 in June 2025 against Alpha’s charter application, citing concerns about its AI-driven instruction, screen time and unverified claims about test score gains. In the months that followed, emails obtained by the reporters show Morath and TEA officials following up with Alpha’s associates and helping coordinate pilots.
The reporting describes pilots in Houston ISD, Fort Davis ISD and Aldine ISD, where roughly 300 students across four middle schools are using the platform for math intervention. At least 10 districts were approached, and five declined. TEA told the reporters there is “no formal partnership between TEA and Alpha Schools,” and that it connected interested district leaders while leaving implementation decisions to them. The reporters also note that TEA has not fully responded to their questions or released all requested records. University of Texas at Austin professor David DeMatthews told them, “It raises questions about whether TEA is serving its public interest and is being transparent and accountable to us.”
A few days earlier, FOX5 Las Vegas reported on a pilot in Nevada’s Nye County School District. The district is testing MagicSchool AI to help teachers with lesson plans, worksheets and grading rubrics, and plans to open controlled student spaces in November where students can get feedback on essays before they submit them. The pilot is grant-funded and runs through December, with teacher feedback collected in February to guide what happens next. Participation is voluntary. District technology director Robert Williams said some educators have ethical objections to AI, and “We don’t want to force them to use something that would compromise their personal ethics.”
I’m not writing to judge whether either tool works. That question matters, and it deserves evidence. This post is about something that comes before the tool: how the decision gets made.
What this reveals about AI adoption in schools
AI didn’t create the question of who decides. It just made the stakes higher and the process more visible.
In Chapter 7 of Human Still Required, I argue that judgment can’t be automated, and it can’t be outsourced upward either. A recommendation from a state agency, a vendor, a funder or a famous school is information. It is not a decision. The district still has to own the choice and explain it.
Wise leaders don’t ask, ‘Is this recommended?’ They ask, ‘Is this right for us?’
Human Still Required
Notice what the Texas reporting says about the five districts that declined. Whatever their reasons, they did the work that leadership requires. They looked at an offer that arrived with the weight of the state behind it and asked whether it fit their students, their schedules and their priorities. That is judgment. It is not resistance.
The harder lesson is about trust. In Chapter 9, I write about how trust is rarely lost in big moments. It erodes quietly, when people sense that decisions are being made somewhere they can’t see. When a tool shows up in a classroom and teachers, parents or board members learn later how it got there, the tool now carries that history with it. Even a good tool can’t outrun a process people don’t trust.
Chapter 15 makes the related point that authority is no longer enough. Leaders can mandate a platform. They can’t mandate belief in it.
Influence now matters more than authority. Presence more than pronouncements. Modeling more than messaging.
Human Still Required
That is why the Nye County approach stands out to me. It is small. It is time-bound. It names the purpose (reducing paperwork for teachers and teaching responsible use to students). It builds in a moment to listen. And it treats a teacher’s objection as something to respect, not something to manage.
In Chapter 11, I write about separating fear from refusal and about why silence is the real problem. Voluntary participation is not a weakness in a pilot. It is how you find out what people actually think.
You don’t need everyone to be enthusiastic. You need them to feel safe enough to engage.
Human Still Required
AI adoption in schools will not fail because a tool was imperfect. It will fail when people stop believing the process behind the tool was honest.
What leaders can do now
- Write down how the decision was made, not just what was decided. Before any AI pilot starts, document who proposed it, who was consulted, what concerns were raised and how they were answered. If you would be uncomfortable seeing that account in a newspaper, the process isn’t finished.
- Treat outside endorsements as input, not permission. When a state agency, vendor or partner brings you a tool, run it through the same local questions you would ask of anything else: What problem does this solve for our students? What would we stop doing? Who bears the cost if it doesn’t work?
- Make participation voluntary in the pilot phase, and mean it. Give teachers a real way to opt out without penalty, and ask those who opt out why. Their reasons are some of the most useful data you will collect.
- Put the listening on the calendar before launch. Nye County scheduled teacher feedback for February. Decide now when you will ask staff, students and families what they experienced, and commit to sharing what you heard.
- Brief your board and families early. People can disagree with a decision and still trust the process. They rarely trust a process they learned about after the fact.
If your leadership team wants a structured way to talk through decisions like this together, the 5-Week Book Study walks through the judgment, trust and fear chapters in order.
Frequently asked questions
Should a district accept an AI tool that the state education agency recommends?
A state recommendation is useful information, but it is not a substitute for local judgment. District leaders should evaluate any tool against their own goals, students and capacity, and be able to explain their decision publicly. In the Texas reporting, five districts that were approached chose not to move forward.
Should teacher participation in an AI pilot be voluntary?
In the early phase, I believe it should be. Voluntary participation surfaces honest feedback, respects teachers who have ethical concerns and builds credibility for whatever the district decides next. Nye County School District made its MagicSchool AI pilot voluntary and scheduled teacher feedback to shape future decisions.
How can school leaders build trust when adopting AI?
Be transparent about how the decision was made, not just what was decided. Name the purpose, set a time limit, build in structured listening and share what you learn. Trust grows when people can see the process and influence it.
The work only leaders can do
Every district is going to be offered AI tools this year, some with impressive backing. The tools will keep changing. The leadership work won’t.
Who was in the room? Who was told? Who got to say no?
Those questions are not bureaucratic details. They are how trust gets built or lost. So here is the question I’d ask you to sit with this week: if someone traced the path of the last AI tool that entered your classrooms, would they find a decision your community could recognize as its own?
Sources
- How Morath’s agency linked Texas schools to Alpha’s AI tool, The Texas Tribune (co-published with ProPublica), October 1, 2026
- Nye County School District pilots AI tool to ease teacher workloads, prepare students, FOX5 Las Vegas, September 29, 2026