Workers Use AI Twice as Often as Businesses Do

Statistics Canada published its first release from a regular series on AI use by Canadian workers on July 30, 2026. In March 2026, 35.9% of workers had used a generative AI tool as part of their main job or business in the previous twelve months. Counting every kind of AI or automation technology, the figure was 41.6%.

Set that beside the other Statistics Canada figure this site argues from. In the second quarter of 2026, 19.2% of Canadian businesses used AI to produce goods or deliver services. Worker use is running at roughly twice business use.

That gap is not a contradiction in the data. It is a description of where the tools actually are: on personal accounts, in personal browsers, chosen by whoever is doing the work, and in most cases outside anything the employer has decided.

What the survey measured

The figures come from supplementary questions on the Labour Force Survey, asked of workers aged 15 to 69 in the provinces, collected in March 2026. That is a large, representative sample rather than a vendor poll, which is why it is worth reading closely.

The awareness numbers are the ones that surprise people:

93.4% of workers were aware of generative AI tools.
51.5% were familiar with how those tools could apply to their own work, including 15.0% who said they were very familiar.
Among workers who had used the tools, 31.4% used them daily and 38.3% a few times a week.

The spread across industries is very wide. Professional, scientific and technical services reported 65.6%, finance and insurance 59.2%, and educational services 53.0%. At the other end, accommodation and food services reported 16.3%, agriculture 17.5%, and trades, transport and equipment operators 14.7%. Management occupations sat at 75.1%.

The number an owner should read twice

Statistics Canada asked the 64.1% of workers who had not used generative AI why not. Most gave the answer this site was built around: 56.0% said it was not applicable to their work. Then, near the bottom of the list:

5.0% said a company or organizational policy limited their use.

Almost nobody is being stopped by a rule, because in most workplaces there is no rule. The tools are already in the building and the decision about how they may be used is being made one desk at a time.

What this changes for an Ontario business

If you already have a written use policy, nothing. This release confirms the scale of something you have accounted for.

If you do not, the honest reading is that your organisation has already adopted AI, without choosing it, and without deciding what may be pasted into it. Two Ontario obligations sit directly on top of that.

The tools see personal information

The Personal Information Protection and Electronic Documents Act reaches what your staff put into a prompt exactly as it reaches what they put into a database, and the Privacy Commissioner of Canada confirmed in May 2026 that Canadian privacy law applies to training data, prompts and outputs. The practical version is in what PIPEDA means for the AI tools your business uses.

Employers of 25 or more who are hiring

Since January 1, 2026 the Employment Standards Act, 2000 and Ontario Regulation 476/24 require Ontario job postings to disclose the use of artificial intelligence in screening, assessing or selecting applicants, for employers with 25 or more employees. A hiring manager using a generative AI tool to shortlist applicants is inside that duty whether or not anybody bought the tool centrally. The detail is in the Ontario AI job posting disclosure rule.

What to do this month

One page, four questions, written down and circulated:

Which tools are approved, and which accounts they may be used from.
What must never be pasted into a tool the business does not control.
Who to ask when the answer is not obvious.
That the person who sends something out stays responsible for it.

That document is the whole job for most businesses of ten to fifty people, and it takes an afternoon. Where it needs more than an afternoon is a business handling health, financial or client identity information, or an employer of 25 or more that is actively hiring. Both cases are covered by AI policy and staff training, and the longer argument for writing it now rather than after an incident is in your staff are already using AI.

None of this is legal advice. Verify anything you intend to rely on against the current statute or with a lawyer.

Questions on this story

Does this mean my staff are using AI without telling me?

Some of them, probably. Statistics Canada did not ask workers whether their employer knew, so nobody can put a number on it honestly. What the survey does show is that 35.9% of workers used generative AI at their main job while only 19.2% of businesses reported using AI to produce goods or deliver services, and that just 5.0% of non-users were held back by a company policy. Worker use running at roughly twice business use, with almost no policies in the way, is what unmanaged use looks like in aggregate.

Is it a problem if staff use ChatGPT at work?

It depends entirely on what they paste into it. Internal notes, public information and draft copy are usually fine. Client names, health details, financial records, employee files and anything else that identifies a person are a different matter, because the Personal Information Protection and Electronic Documents Act applies to that information in a prompt exactly as it applies to it in a database. The distinction is easy to write down and impossible to guess, which is why the written rule matters more than the tool choice.

We are a twelve person company. Do we really need a written AI policy?

Yes, and it is a short document rather than a project. One page covering which tools are approved, what may never be pasted into a tool the company does not control, who to ask when the answer is unclear, and the rule that a person stays responsible for anything they send out under their own name. Twelve people generate twelve different private judgements about that boundary if nobody writes one down.

Which industries in the survey used generative AI the most?

Professional, scientific and technical services led at 65.6%, followed by finance, insurance and real estate at 59.2% and educational services at 53.0%. Accommodation and food services came in at 16.3% and agriculture at 17.5%. By occupation the spread is wider still: 75.1% in management against 14.7% among trades, transport and equipment operators.

Before you spend anything

Tell me how many people work there, what the busiest hour of the week looks like, and which task everybody complains about. That is usually enough to say on a first call whether an assessment is worth your $999 or whether you have one obvious problem that needs one obvious fix.

Reply within one business day · Mon to Fri, 9am to 5pm Eastern