The numbers are striking. Fifty-five per cent of Americans now believe artificial intelligence will do more harm than good in their daily lives — an 11-point increase in just twelve months. Seventy per cent think AI will reduce job opportunities. Eighty per cent say they are concerned about the technology. And yet more than half are already using it to research topics, a quarter are deploying it for workplace projects, and nearly a third of employed people are using it on the job right now.

This is not a picture of a workforce rejecting AI. It is a picture of a workforce adopting AI with deep unease — and for Canadian HR leaders, that distinction matters enormously.

The findings come from a new Quinnipiac University national poll of 1,397 U.S. adults published yesterday. While the survey reflects American attitudes, the underlying dynamics — accelerating adoption, eroding trust, sharpening anxiety about job security — map closely onto trends playing out in Canadian workplaces, and onto the productivity challenge that Statistics Canada has been documenting for years. 

The trust gap is the real story

The headline numbers on AI sentiment are eye-catching, but the most strategically significant finding for HR leaders lies in the contradiction between use and trust.

Read next: Uber staff are using an AI clone of their CEO to practice pitches

Fifty-one per cent of respondents say they use AI for research. Twenty-seven per cent use it for school or work projects. Twenty-seven per cent use it to analyse data. But only 21 per cent say they trust AI-generated information most or almost all of the time. Seventy-six per cent say they can trust it only some of the time, or hardly ever.

In other words, nearly one in three working Americans is regularly using a tool they fundamentally do not trust to produce reliable outputs. In a Canadian workforce context, that should concern HR executives on two levels: first, because untrusted tools tend to be used inconsistently and uncritically; and second, because organisations that deploy AI without addressing the trust deficit risk amplifying the anxiety that is already driving resistance. 

Gen Z is the most worried — and the most capable

Perhaps the most counterintuitive finding in the Quinnipiac data is that younger workers — those with the greatest familiarity with AI tools — are simultaneously the most pessimistic about what the technology will mean for their careers.

Among Gen Z respondents (born 1997–2008), 81 per cent believe AI will lead to a decrease in job opportunities — the highest figure of any generation, and 15 points above baby boomers. Gen Z is also the generation with the highest reported AI usage and the strongest technical fluency.

As one of the Quinnipiac researchers noted, AI fluency and labour market optimism are moving in opposite directions among younger workers. For HR leaders in Canada managing talent pipelines, graduate recruitment, and early-career development programs, this finding demands attention. The workers most likely to drive your AI transformation agenda are also the most likely to be questioning whether they have a future in the organisation on the other side of it.

Read next: What's behind Canada's weak productivity growth?

The implications for retention, psychological safety, and career pathing are significant. Organisations that can give younger employees genuine visibility into how roles are evolving — rather than offering vague reassurances — are likely to maintain a competitive advantage in attracting and keeping this cohort. 

Job displacement fears are rising — but concentrated differently than expected

The data challenges a common assumption: that blue-collar workers are more threatened by AI than their white-collar counterparts. In fact, the Quinnipiac poll found that 73 per cent of blue-collar workers and 71 per cent of white-collar workers believe AI will decrease job opportunities — a near-identical split.

This matters for Canadian HR professionals in industries that have historically framed AI as a threat to manual and administrative roles while positioning knowledge workers as beneficiaries. The workforce is not making that distinction. Concern is broadly distributed, and any communications or change management strategy built on the assumption that professional employees feel insulated is likely to miss the mark.

Read next: AI adoption key to job security: report

What does differ is current AI adoption at work: 49 per cent of white-collar workers report using AI on the job, compared with 18 per cent of blue-collar workers. The adoption gap suggests that white-collar employees may be grappling more immediately with the practical reality of AI integration — and the anxiety that comes with it. 

The AI supervisor question: a red line for most workers

One finding in the Quinnipiac poll stands out as a clear signal to HR leaders designing workforce AI programs: 80 per cent of Americans say they would be unwilling to have a job where their direct supervisor was an AI program that assigned tasks and set schedules.

This is not simply a preference. It reflects something deeper about how workers understand authority, accountability, and the employment relationship. The desire for human oversight — even where AI may be more accurate or efficient — runs throughout the data. When asked whether they would prefer AI or a human to read their medical scans, even if the AI were proven more accurate, 81 per cent chose a combination of both. Only three per cent said AI alone.

Read next: The career ladder is disappearing because of AI. Are taxes making it worse?

For Canadian organisations deploying AI-assisted performance management, scheduling, workload allocation, or productivity monitoring tools, the message is clear: the technology may be capable, but worker acceptance depends on maintaining meaningful human involvement in decisions that affect people's working lives. Removing the human from that loop — even partially — is likely to generate resistance that undermines the productivity gains the organisation is seeking. 

The Canadian productivity connection

The Quinnipiac findings land at a moment when Canada is confronting an uncomfortable productivity reality. A new Statistics Canada study, published this week, confirms what many economists have been warning: weak competitive pressure is dragging down labour productivity growth. The research found that frontier firms — those in the top tenth percentile of productivity — actually suffer the sharpest declines when competitive intensity softens, as they ease off on the innovation and investment that drove their lead.

AI is widely identified as one of the most powerful levers available to address that productivity deficit. SAP's 2025 Future of Work research found workers are saving an average of 75 minutes a day through AI, and EY-Parthenon estimates AI could lift economy-wide labour productivity by 1.5 to three per cent over the next decade, with the largest gains concentrated in financial services, professional services, consulting, and legal — precisely the sectors that dominate Canada's white-collar employment base.

Read next: What the world's largest study of work hours reveals about how we work

But the Quinnipiac data suggests the productivity dividend will not arrive automatically. It will require organisations to do the harder work of building the trust, capability, and cultural conditions under which workers actually use AI well — consistently, critically, and confidently — rather than reluctantly and inconsistently. 

What senior HR leaders should do now

The evidence across these data points points to several priorities for Canadian HR executives.

The first is transparency. Seventy-six per cent of Americans say businesses are not doing enough to be transparent about how they use AI. In a Canadian context where privacy expectations are high and employee relations are governed by a web of provincial and federal obligations, that transparency deficit represents both a legal exposure and a talent risk. Organisations that communicate clearly about where AI is being used, how decisions are made, and what human oversight exists are likely to fare significantly better in both regulatory scrutiny and employee trust.

The second is career visibility. The data on Gen Z pessimism underscores the urgency of giving employees — particularly early-career staff — concrete, credible answers to the question: what does my career look like in an AI-augmented organisation? Vague promises about "humans and AI working together" are no longer sufficient. Workers want specificity about roles, skills, and pathways.

The third is change management that acknowledges anxiety rather than dismissing it. Eighty per cent of Americans express concern about AI — and that concern is consistent across every age group and income level in the Quinnipiac data. HR communications strategies that lead with productivity benefits while glossing over legitimate workforce concerns are likely to increase resistance, not reduce it.

The fourth — and perhaps most fundamental — is ensuring human oversight remains visible and genuine in AI-assisted people processes. Performance management, scheduling, hiring, and workforce planning are areas where AI tools are being deployed at speed. The data suggests workers will accept AI as a tool, but not as a decision-maker. The organisations that get this right will build the trust that unlocks real productivity gains. Those that get it wrong will face a workforce that is simultaneously using AI and fundamentally uncertain about its place in the organisation that deployed it.