In most industries, AI adoption is framed as a race.
But Education operates under a different set of pressures.
Scott Mckenzie has spent the past six years reshaping how Fort Vermilion School Division collects and validates student data. The goal was not simply to implement AI, first he needed to achieve clarity on the mission. AI entered the conversation later.
McKenzie says that Canadian K–12 education is still early in its AI journey. And while the important conversations about how to best use AI in education have begun, actual deployments remain limited.
In fact, much of the early work is raising literacy to help people understand the best tools for the job.
“There’s still a disconnect between what people believe AI is and what AI isn’t,” Mckenzie says. In schools, the fear is often about autonomy. Teachers worry that decisions may be outsourced to a model. Parents worry about how student data is used.
With so much at stake, building trust among all groups is essential.
Trust moves more slowly than technology
When enterprises adopt AI, they do so to improve business performance. Leaders in education must balance innovation with public scrutiny, constrained budgets, and strict regulatory oversight.
While both sectors pursue transformation, the context in which they deploy AI fundamentally shapes their idea of what success looks like.
“When your stakeholders are parents rather than shareholders, experimentation looks very different,” Mckenzie explains.
Before embarking on any AI initiatives, his focus was on building the necessary foundations.
“You need to have clean data. You need to have governance. You need to have collection methodologies,” he says. “If you don’t, AI is always just going to be what it believes it is.”
That line captures the heart of his position. Weak foundations will not support strong AI.
“Once you get into AI, you chase the AI. You don’t chase the data behind it. Nine times out of 10, it’s the details behind it that need fixing.”
For senior data leaders, the warning is familiar. The temptation to launch pilots can outpace the discipline to define metrics. In education, that discipline is not optional. It underpins public trust.
Literacy is now about empowering judgment
AI has the potential to offer powerful insights to teachers without requiring extensive technical expertise.
But it also introduces new risks.
Mckenzie points to a familiar pattern. Challenge an AI response and it may revise its answer. Challenge it again and it may revise it once more. The fluidity can be helpful, but it also reveals something important.
“If AI changes its answer when you question it, your workforce needs to be able to question it too.”
In this environment, data literacy must be more than knowing how to build a dashboard. Users need to know how to recognize when an output does not align with reality.
Over time, that feedback strengthens the system, while without it, the organization begins to drift.
The talent constraint is real
Even with clear governance and strong literacy, AI capabilities in public education lags due to competition for talent with the private sector where professionals can expect more money and fewer constraints.
What’s more, finding someone who understands machine learning is hard enough. Finding someone who understands it in the context of student assessment and provincial reporting requirements is harder still.
External vendors can close gaps quickly. They bring expertise and speed. They can deploy tools in weeks rather than months. But external reliance too, carries risk.
Building internally offers continuity and alignment with organizational values. It also takes time and patience. Most school systems, and many enterprises, will land somewhere in the middle.
A sector that shapes the future workforce
There is another layer to education’s cautious approach.
While debating how to deploy AI responsibly, schools are also preparing the workforce that will lead AI transformation elsewhere. The same classrooms that hesitate to automate assessment are shaping future data scientists, engineers, and executives.
Mckenzie takes that dual responsibility seriously.
Education does not exist to optimize transactions. It exists to develop judgment. That mission changes how technology is evaluated. Efficiency is important. Equity, privacy, and clarity matter just as much.
As Mckenzie prepares for CDAO Canada, he is less concerned with declaring success and more interested in dialogue.
“What do you deem success? What are your measures? What are your metrics?” he asks. Even within one sector, definitions vary. Across sectors, they vary more.
The common ground lies in discipline. Clear guardrails. Honest conversations about risk. A willingness to learn from peers in different industries.
“Nobody’s right, nobody’s wrong,” he says. “It’s what works for you at the moment. But that doesn’t mean you can’t take away ideas.”
In an era that celebrates speed, education offers a counterpoint. Move carefully. Define terms. Build trust first. Then scale.
For senior data leaders, that may be the most provocative lesson of all.
__
To explore how leaders at Canada's biggest data, analytics, and AI are building trusted, human-centered AI strategies join us CDAO Toronto on March 25th.