Canadian workers feel guilty about using AI at work

Employment Hero has published research showing many Canadian workers feel guilty about using artificial intelligence at work. The report found that 43% feel guilty when they use AI to produce work.

The findings highlight workplace tension as AI becomes more common in office jobs but remains poorly understood by many employees and managers. The survey found that 39% of Canadian workers believe using AI for parts of their job feels like cheating, while 34% hide their use of AI from their employer.

Among younger staff, the discomfort appears stronger. More than half of Gen Z workers, 56%, said they feel guilty using AI at work.

The study suggests hesitation over AI is not simply a matter of willingness to use the technology. Only 41% of Canadian workers believe their AI skills are sufficient for an AI-driven labour market, while 60% rate their competence as low to average.

Training appears to be a major gap. Just over half of respondents, 51%, said their employer does little or nothing to develop AI skills, leaving many workers to learn elsewhere. A majority, 58%, said they had picked up AI skills through social media.

That pattern raises questions for employers about oversight as well as education. The report found that 45% of businesses believe employees are using personal AI accounts at work, a sign of so-called shadow AI, in which staff turn to tools outside approved systems or company policies.

Confidence gap

The results come as employers face pressure to decide how AI should fit into day-to-day work. For many organisations, the issue is moving beyond access to tools to setting rules on when AI can be used, what data can be entered into systems and how staff should check outputs before sharing them.

Workers appear to be operating in an uncertain environment. If employees think AI use could be viewed as dishonest, they may be less likely to disclose when they have used it for drafting, research, summarising information or other routine tasks.

That uncertainty can also affect risk management. Employees using consumer AI products through personal accounts may expose company information to services that have not been approved internally, especially if businesses have not set clear boundaries on what can and cannot be shared.

Chris Pinkerton, Managing Director at Employment Hero Canada, said the figures show a broader issue around workplace culture and guidance.

"This research shows that Canada's challenge isn't AI adoption - it's AI confidence," said Pinkerton. "Workers already recognize AI is becoming an essential workplace skill, but many still feel they need to hide using it because they don't have clear guidance or confidence in what's acceptable. The organizations that will succeed won't simply be the ones adopting AI fastest--they'll be the ones that create a culture where employees feel empowered to use it responsibly, transparently, and confidently."

Skills and policy

The survey covered 3,290 business leaders and 5,454 workers across Australia, Canada, New Zealand and the UK. In Canada, the sample included 1,001 business leaders and 1,500 workers.

Although the report focuses on employee sentiment, the findings also underline the management task facing businesses. If AI use is becoming routine while internal policy remains vague, employers may struggle to balance productivity gains with compliance, confidentiality and quality control.

Many of the concerns identified in the research centre on basic governance rather than advanced technical deployment. Clear policies on approved tools, disclosure of AI assistance and the role of human review may do as much to reduce anxiety as formal training programmes.

For employers, the practical challenge is cultural as much as operational. Staff who believe AI use is a shortcut or a form of cheating may avoid discussing it openly, making it harder for managers to understand how widely such tools are being used and for what tasks.

The data also suggests a split between the speed of AI adoption and the pace of workplace adaptation. Employees are already experimenting with tools on their own, but many companies appear not to have kept up with training, internal communication or skills development.

That leaves workers filling the gap themselves, often through informal channels. Social media can offer quick tips and examples, but it is not a substitute for company-specific guidance on confidentiality, accuracy standards or the limits of acceptable use in regulated or sensitive work.

As businesses assess how to integrate AI into normal workflows, the report suggests one immediate issue is not resistance to the technology itself but discomfort about using it in view of colleagues and employers. In Canada, 34% of workers said they hide their AI use from their employer.