In the first article in this series, we met the students at The Study, a Montreal girls' school that built its own private AI platform and then watched its students build on top of it. That article told the story. This one asks the question: what should everyone else do?
The Ban That Won't Work
In April 2026, Canada's federal Liberal Party voted to back a minimum age of sixteen for social media accounts and AI chatbots.[1] Manitoba said it would be the first province to enforce it.[2] A national poll found that more than two-thirds of Canadians agreed.[3]
The worry is real. Children are using powerful tools that most adults do not fully understand. The impulse to protect them is natural.
But the plan does not hold up.
Michael Geist, a University of Ottawa law professor and one of Canada's leading voices on digital policy, called the push "the illusion of protection."[4] A ban, he argued, would not stop young people from using AI. It would push them toward tools built in other countries – tools with no safety filters, no content rules, and no way to call for help. The technology to enforce a ban without invading everyone's privacy does not exist yet.
Meanwhile, the students are already there. A 2025 survey found that 73 percent of Canadian students use AI for schoolwork, up from 52 percent two years earlier.[5] Daily use jumped 15 percent in one year alone. Teachers are starting to come around too. Some told Policy Options that ignoring AI is more dangerous than using it.[6]
Banning AI for young people is like banning the tide.
The question is not whether students will use AI. It is whether anyone will teach them to use it well.
– Isabelle Plante
Tool Fluency, Not Tool Loyalty

Jean-Baptiste-Siméon Chardin, "The Schoolmistress" (c. 1735). National Gallery, London. The tools are simple. The method is what matters. Public domain
At The Study, students do not just use AI. They build with it. They pick their own tools – Lovable, Claude Code, OpenAI Codex, Rosie (the school's AI platform) – and the school does not tell them which one to choose. "Because when they leave The Study, there's going to be no one dictating to them," said Amalia Liogas, the school's Director of IT. "They have to understand why to choose what, and for what reason. If you haven't tested it, you don't know."
The school's AI Framework calls this design thinking: students "iterate, experiment, and learn through failure within a structured and supportive environment."[7] The tools change every year. The method does not.
This is what we call tool fluency. It does not mean loyalty to one platform. It means the ability to pick a tool, learn it, adapt it, or build your own – quickly and on the merits. The students using AI this year will use a different tool next year. The class after them will replace it again. Some will build their own.
For policymakers, the difference matters. Banning treats AI as a threat to manage. Tool fluency treats it as a skill. The difference shows up after graduation, when nobody is managing the ban anymore.
Where Other Schools Are
Most schools are still in the research phase. Ms. Liogas sees it up close. "I think it's scary to make that leap," she said. "If you don't have confidence, you're still in the research phase."
She does not pretend it has been easy. "Am I saying things have been 100 percent peachy keen with us? Of course not – but that's how you learn." When she started teaching vibe coding in January, she told her students: "I don't know vibe coding, but we're going to learn together and make mistakes together. Being honest with them is the best policy while figuring this out."
That honesty matters. Schools do not need to have all the answers before they start. They need a method and the willingness to learn alongside their students.
What You Can Do

Judith Leyster, "Self-Portrait" (c. 1630). National Gallery of Art, Washington. She chose her own brush, her own canvas, her own subject. CC0
Here are four moves for any school, board, or parent who wants to begin.
Own your foundation. Free everything above it. A school's AI plan has two layers. The bottom layer – the AI platform, the data rules, the infrastructure – should belong to the school. Everything above it – the tools students pick for their own projects, the things they build – should be open. The job is to build a floor strong enough that students can stand on it and build anything.
Let students build, with the tools they pick. The best learning happens when students create something real. The Imagination Age does not reward students who memorized the right answer. It rewards students who can ask a better question.
Start with the method. Tools without a method lead to chaos – in classrooms and in companies. An AI consultant in Canada recently walked away from a three-year client after watching the CEO open seven coding windows at once, push untested changes to a live product, and break systems the consultant had spent years building.[8] The tools were the same ones the consultant used every day. The difference was the method. The Study understood this from the start: Rosie is not just an AI. She is the first step in a structured process – what manufacturing calls poka-yoke, or mistake-proofing.[9] Teacher review, quality checks, and human oversight are built into the system, not added later. The safe path is also the easy path.
Make girls' voices the default, not the exception. The Information Age was built mostly by men. The Imagination Age is being built right now. The next decade of AI will be shaped by whoever is in the room when it is made. Whatever school you support, whatever board you sit on – the question is not whether girls are interested in AI. They are. The question is whether you have built a place where they can imagine with it.
Ms. Liogas's advice to anyone who wants to start is simple. "Your 'why' is really important," she said. "Are you doing this because it's a checkbox, or because you want kids to learn? If you want kids to learn, that's your answer right there. Your focus is going to be on the pedagogy: understand how to incorporate it into classes, how kids can use it, and what it means to their futures."
About Sage.Education
Rosie runs on Sage.is AI UI, the platform that powers Sage.Education and Sage.is, a Canadian AI platform that lets schools and small businesses run their own AI on their own hardware. The Study already believed in ownership, transparency, and student freedom. Sage.is AI UI was the platform that matched those values.
Footnotes
Disclosure: Rosie runs on infrastructure built by Sage.is. This article is part of the Sage.Education series. The editorial decisions, reporting, and final prose are the author's. Quotes from Amalia Liogas were captured in a May 2026 interview and used with permission.
Canadian federal Liberals voted at their April 2026 policy convention to back a minimum age of 16 for social media accounts and AI chatbots. Culture Minister Marc Miller said the government was "seriously" considering a ban or moratorium. Bloomberg ↩︎
Manitoba announced it would be the first Canadian jurisdiction to ban youth under 16 from using social media and AI chatbots, starting in schools. CBC News ↩︎
More than two-thirds of Canadians support banning access to social media and AI chatbots for children under 16. May 2026 poll. National Observer ↩︎
Michael Geist, University of Ottawa law professor. "The Illusion of Protection: Why Canada's Growing Push to Ban Social Media for Kids Won't Work," April 2026. michaelgeist.ca See also: "Government Has a Choice: Why an AI Chatbot Ban for Kids is an Even Worse Idea Than a Social Media Ban," May 2026. michaelgeist.ca ↩︎
KPMG survey, October 2025. 73% of Canadian students rely on generative AI for schoolwork, up from 52% two years earlier. Daily use jumped 15% in one year. KPMG Canada ↩︎
"How Canada's AI patchwork is failing students in the classroom," Policy Options (IRPP), January 2026. policyoptions.irpp.org ↩︎
The Study, AI School Framework (internal document, shared by Amalia Liogas, Director of IT). thestudy.qc.ca ↩︎
"Seven Windows and No Method," Sage.is, May 2026. Sage.is/resources ↩︎
Poka-yoke (Japanese: "mistake-proofing") is a method from the Toyota Production System. It means designing a process so that mistakes are caught before they cause damage, or made impossible to make. In schools, this looks like teacher review, quality checks, and structured steps built into the system – so the safe path is also the easy path. Lean Blog ↩︎
Sage.Education