NetworkChuck took AI
away from his kids.
Here is why he is right — and what to do instead.

NetworkChuck — real name Chuck Keith, 5.3 million YouTube subscribers, one of the most trusted voices in networking, cybersecurity, and technology education — posted something on LinkedIn that most people in his position would not say publicly.

The post has 2,951 reactions, 328 comments, and 128 reposts. For a LinkedIn post about a parenting decision, that is a significant response — because it touched something that a lot of people in tech, people who build AI systems and use them every day, are privately wrestling with about their own children.

Read the post carefully. He did not just take AI away. He gave his children a structured alternative: physical books. A self-hosted, searchable Wikipedia instance — local, no AI summaries, no Google. And writing practice with no AI assistance. Not even spell check.

That is not technophobia. That is a deliberate, thoughtful architecture for cognitive development. A person who understands AI deeply enough to run local models at home for his children decided that the right thing for those children right now is to develop the capacity to think before they are given a tool that can think for them. That distinction — think before you prompt — is the entire argument in five words.

The numbers confirm what NetworkChuck felt

Global student AI usage jumped from 66% in 2024 to 92% in 2025 — a 26 percentage point increase in a single year. That is not gradual adoption. That is a near-complete shift in how students approach work that requires thinking.

By the start of 2026, 86% of students in higher education are using AI as their primary research and brainstorming tool. For high schoolers, 80% are receiving some form of formal AI literacy education. But only 8% of students in Pre-K through 3rd grade are receiving comparable training. The youngest, most cognitively vulnerable children are the least prepared and the least protected — and the most exposed.

64% of US teenagers say they use AI chatbots regularly. 72% of American teenagers have used AI chatbots as companions. Not just for homework. As companions. That is a different conversation from "my child uses ChatGPT to help with essays."

New York City just did what NetworkChuck did — at scale

On September 2, 2026 — six days ago — New York City Mayor Zohran Mamdani announced a one-year moratorium on student-facing generative AI for children in grades 2-K through 8th grade. The ban covers nearly 600,000 students in the country's largest public school system and bans companion chatbots across every grade.

Mamdani's statement at the press conference is the clearest articulation of the NetworkChuck position applied at policy scale:

That last sentence is the sharp edge of the argument. The evidence base for AI being beneficial to young learners is almost entirely funded by the companies selling the AI. Independent research on cognitive development and AI dependency in children is still emerging — and what exists is not encouraging.

Mamdani also announced the ban alongside Schools Chancellor Kamar Samuels. City Hall is calling it the most expansive student-facing AI policy in the country. The moratorium is one year — enough time to develop an evidence-based framework rather than letting the technology shape the children before anyone has measured what it is doing to them.

The cognitive development argument — why this matters at a brain level

The case for limiting AI access for young children is not nostalgic. It is neurological.

Cognitive development in children depends on the experience of working through difficulty. When a child sits with a problem they cannot immediately solve — a maths question, a paragraph they need to write, a concept they do not understand — and works through it, they are not just learning the answer. They are building the neural architecture for independent reasoning. The struggle is not a bug in the learning process. It is the process.

AI removes that friction completely. It does not help a child work through a problem — it removes the problem. The answer arrives before the thinking begins. For an adult who already has the cognitive infrastructure — who can evaluate the AI's output, identify its errors, use it as a starting point rather than an endpoint — that is a productivity tool. For a child whose cognitive infrastructure is still being built, it is a shortcut around the construction site.

The analogy that holds up: calculators in primary school mathematics. Most educators agree that children should learn to do arithmetic manually before they are given calculators — not because calculators are bad, but because understanding what a calculation means requires experiencing it. A child who has never done long division does not understand what division is — they understand what buttons to press. The same principle applies to AI at a much broader cognitive scale.

The legislation catching up — 134 AI bills in 2026

134 AI-in-education bills have been introduced across 31 states in 2026. That is not a fringe concern — it is a legislative consensus forming in real time that something needs to govern how AI interacts with developing minds.

In March 2026, Senator Markey introduced legislation that would prohibit AI chatbots from displaying advertisements to minors, ban companies from training models on minors' personal data, and prohibit profiling based on minors' AI chatbot inputs. The legislation frames AI access for children as a data and safety issue as much as a cognitive one.

Character.AI closed off open-ended chat for users under 18 in late 2025. A platform built on AI companionship decided, under regulatory pressure and public scrutiny, that its product was not appropriate for minors in its unrestricted form. That is a significant admission from a company whose entire business model depends on users forming ongoing relationships with AI characters.

What to do instead — empowering kids without letting AI think for them

The answer is not to pretend AI does not exist or to raise children who are afraid of it. NetworkChuck did not delete the internet from his house. He removed a specific tool from children whose minds were not yet ready to use it without substituting it for their own thinking.

The goal is children who can think independently first and use AI as a multiplier second. That sequencing is everything. A child who has learned to reason, to write, to solve problems, to sit with confusion and push through it — that child, given access to AI at the right developmental stage, becomes genuinely more capable. A child who has never developed those skills and is given AI instead simply becomes dependent on it.

The honest position — AI is coming regardless

NetworkChuck knows AI is not going away. He is not raising children who will never encounter it — he is raising children who will encounter it with a developed mind rather than a dependent one. There is a significant difference between a 22-year-old who grew up using AI as a crutch and a 22-year-old who grew up thinking independently and then learned to use AI as a lever.

The first person can produce AI output. The second person can produce AI-augmented thinking. Those are not the same thing, and the gap between them will matter more and more as AI becomes ubiquitous. The people who will have genuine competitive advantage in an AI-saturated world are not the ones who started using AI earliest. They are the ones who can think clearly, evaluate critically, and use AI to amplify reasoning they already possess.

Mamdani framed New York City's policy as follows: "In the age of AI, we must prepare our children for the future — to be AI creators, not just consumers." That is the right frame. A consumer of AI outputs is a person who relies on what AI produces. A creator — someone who understands systems, who can reason independently, who can direct and evaluate AI rather than simply receive it — is a person who builds things.

NetworkChuck has spent his career teaching people to build things. His decision about his own children is entirely consistent with that. You learn to think first. Then you learn to use the tools that amplify thinking. In that order.