A parent and school board member recently asked me how many hours her child spends on a Chromebook during the school day. It is a fair question, and one most district technology leaders field regularly.
I gave her a number estimate based on what I could gather from the data reporting. Then I told her what I have come to believe is the more useful question: what cognitive work is her child doing during that time?
CoSN's 2025 Blaschke Report and Screen Time Toolkit is helpful because of its refusal to treat "screen time" as a single thing. Cooper Sved's three-category framing of cell phones and social media, educational technology, and at-home entertainment is the right starting point, and the Screen Time Continuum gives families and teachers a vocabulary they can actually use. The toolkit moves the conversation from quantity toward quality and context, which is exactly where it needs to go.
That shift matters.
But generative AI now extends this matter even further. A student scrolling social media, watching entertainment videos, completing an adaptive math lesson, collaborating in a shared document, and using a chatbot to draft an essay are all technically using screens. They are not having the same experience. They are not practicing the same habits of mind. They are not encountering the same risks. The harder question for CTOs, technology and innovation directors, instructional technology teams, superintendents, and the principals making AI calls in real time is not how much screen time is too much. It is what the screen is asking students to practice.
The harder question is not how much screen time is too much. It is what the screen is asking students to practice.
Generative AI complicates this in a specific way. A student may be using an approved device, inside a district-managed platform, on a sanctioned assignment, while still bypassing the very thinking the task was designed to develop. The minutes still log. The interface looks identical to anyone walking past. Most monitoring tools cannot tell the difference at the level that matters. But cognitively, those students may be doing very different work.
This is the central problem my doctoral research examines and many other amazing leaders and practitioners are also grappling with. Generative AI does not simply add another tool to the EdTech ecosystem. It changes the conditions under which students produce academic work. A student can now generate fluent writing in multiple languages, plausible explanations, polished summaries, code, and arguments without doing the interpretive work that usually produces understanding.
Learning scientists have a name for this. Manu Kapur calls it unproductive success — correct-looking performance without durable learning. The output is right. The schema is not built. In school terms, this means a student can submit something that looks successful while missing the struggle, revision, uncertainty, dialogue, and ownership that the assignment was meant to cultivate.
A student can now generate fluent writing without doing the interpretive work that usually produces understanding. Manu Kapur calls this unproductive success.
This is not only an academic integrity problem. It is a learning design problem.
Districts already know how to write policies that prohibit cheating, block certain tools, approve others, vet vendors, and craft acceptable-use language. Those steps matter. Privacy, security, compliance, accessibility, and procurement are essential responsibilities for any school system. They are also not sufficient. The deeper question is whether our instructional systems preserve the kinds of human learning processes that AI can now make optional.
I use the term pedagogical friction to describe the deliberately designed moments of productive struggle that learning requires. It is the deliberate protection of the thinking, dialogue, authorship, and institutional support that help students build understanding rather than only produce outputs.
For practitioners, pedagogical friction can be understood through four dimensions. Select each to explore the full framework.
The equity dimension belongs in the same conversation. Not all friction is productive. The goal is not to preserve all difficulty — it is to distinguish what builds capacity from what simply blocks access.
Productive Friction
Difficulty that builds capacity — the cognitive and social work that produces durable understanding.
Exclusionary Friction
Barriers that block access without benefit — difficulty that gatekeeps rather than develops.
For multilingual learners, students with disabilities, and students with limited academic support at home, technology can reduce exclusionary barriers. Text-to-speech, translation support, vocabulary scaffolds, sentence starters, and AI-supported feedback may open doors that were previously closed. The goal is not to preserve all difficulty. The goal is to distinguish productive friction from exclusionary friction.
That distinction reshapes the district question. The useful question is not "Should students use AI?" It is "At what point in the learning process should AI enter, and what human thinking should happen first?"
The difference between "AI instead of thinking" and "AI after thinking" may become one of the most important instructional design decisions of the next decade.
For district leaders, this shift has practical consequences. Select each priority to mark progress on the leadership agenda.
The CETL Framework of Essential Skills already names instructional focus, leadership and vision, and ethical and legal judgment as core to our work. Generative AI pulls all three into the same conversation. The question of what cognitive work we protect for students — and what work we are willing to delegate to algorithmic systems — is a policy question, an instructional question, and a values question at once. It is not a question we can leave to vendors, and it is not a question a screen time policy by itself can answer.
The Blaschke Report opened the door to a more honest conversation about technology in our schools. Generative AI is now standing in that doorway. Some screen use is creative. Some is extractive. Some helps students think. Some helps them avoid thinking. Our job as district technology leaders is to keep the friction that learning needs and remove the friction that learning does not.