Access to AI tools is becoming easier. Knowing when to use them, how to evaluate their output and where responsibility remains with a person is a different capability. A new Canadian announcement puts that distinction on the business agenda.
On 9 September 2026, the federal government launched the National AI Literacy Initiative with the Alberta Machine Intelligence Institute, Amii. It announced free learning streams for students, educators and Canadians. The rollout is phased: additional post-secondary institutions can join the student consortium from 21 September, while the first complete version of the Canadians stream is planned through community partners later this year. The announcement also directs workers and job seekers to short-duration training through Job Bank's Training Finder. Source: Government of Canada, 9 September 2026.
The business question behind AI literacy
CREDIUM's view is that workforce readiness should be defined through observable work. A team member who completes a course has gained exposure to the material. A team member who can recognise an unsupported statement, protect restricted information and explain a recommendation has demonstrated something the organisation can use.
That distinction matters when managers allocate learning time. General awareness creates a foundation, but each business still needs to connect learning with its own tools, information and standards. The most useful training plan starts with the decisions employees are already expected to make.
Define the skills by role
A researcher, a client-facing employee and a manager may use the same AI application while needing different judgement. Build a short role map before selecting courses or purchasing more licences.
- People preparing work: Can they select appropriate inputs, explain the task and identify claims that require verification?
- People reviewing work: Can they compare the output with the source material and identify an omission that changes the conclusion?
- People approving decisions: Can they assess the business consequences and determine when additional expertise is needed?
- People maintaining the process: Can they keep instructions and approved information current as tools and requirements change?
For each role, define one practical demonstration. The objective might be producing an accurate briefing with traceable sources or identifying why a proposed response should not be sent. This gives the learning programme a clear finish line.
Teach judgement with a realistic example
Consider a fictional team preparing a market update for a management meeting. Its exercise includes a company announcement, an older market report and a plausible but unsupported AI-generated statistic. The task is to produce a short briefing that distinguishes what happened, what remains uncertain and what the team thinks it means.
A strong response would identify the unsupported statistic, notice the report's date and avoid describing an announced product as already widely deployed. A fluent summary that misses those distinctions would need more work. This is an illustrative learning exercise, not a reported CREDIUM engagement.
Reviewing the exercise together also reveals whether the organisation's standards are clear. If managers disagree about acceptable evidence, that is a process issue worth resolving before expecting employees to apply a consistent rule.
Give managers their own learning agenda
Managers need to understand how AI changes the work they receive. If drafting becomes faster, review capacity may become more important. If employees generate more options, the team needs a method for comparing them. If a tool begins taking actions, the approval boundary needs attention.
A management session should therefore examine workload, decision authority and the quality of accepted outputs. Asking only how many employees use AI can miss whether the business is getting better work or simply more material to review.
Connect these decisions to an AI governance framework when tools can change records or interact with external systems. Training and operating permissions should reflect the same expectations.
Use a 30-day learning cycle
In the first week, select one recurring task and record the team's current approach. In the second, introduce focused learning and a controlled exercise using approved material. In the third, apply the method to a small set of real tasks with review. In the fourth, discuss the results and update the operating guide.
Measure the quality of accepted work, review effort and recurring mistakes. Keep course completion as an activity measure, alongside evidence of what people can now do. Give employees a clear route to raise uncertainty without feeling they must make every AI output usable.
Turn learning into an operating capability
The national initiative creates a timely opening for a workforce discussion. Employers can use that momentum to make learning specific, proportionate and connected to their actual business priorities.
Start with one role, one task and one practical demonstration. Then connect the findings to the strategy implementation roadmap. For the wider adoption decision, see AI adoption in Canada. The valuable outcome is a team that can use new tools with better judgement and explain the work it delivers.