One in four workers say they plan to leave their current employer because of how that employer views or uses artificial intelligence. That figure, 25.1%, comes from ZipRecruiter's 2026 AI Worker Report, released on 26 October 2026 and based on an online survey of more than 1,500 US workers conducted between 8 and 21 September 2026. The survey covers US respondents, so Canadian People leaders should treat it as a directional signal rather than a domestic benchmark, testing its implications against internal engagement data and exit interview trends specific to their own organisations.

The headline number is striking, but the finding with greater operational weight for HR sits beneath it. Access to AI training fractures sharply by seniority: 83.5% of senior leaders report having it, compared with 46.4% of individual contributors and 45.3% of entry-level workers. The employees whose day-to-day tasks AI is absorbing most rapidly are the least likely to receive structured instruction in how to work alongside it. For HR Directors and People leaders at Canadian organisations, that combination represents a retention risk that is quiet in the current labour market and likely to surface when hiring conditions shift.

Statistics Canada's Labour Force Survey has documented a cooling but still-tight labour market through much of 2026, with voluntary separation rates remaining below their 2022 peak. That context mirrors the US report's central caution: suppressed turnover does not equal satisfied employees, and the gap between the two tends to close quickly when hiring picks up.

What the data tells HR about workforce sentiment toward AI

The ZipRecruiter report does not describe a workforce that rejects AI outright. A majority, 52.2%, say AI has made their work more enjoyable, a share that climbs to 76.5% among senior leaders. Research is the most common reported use, cited by 52.1% of respondents, followed by summarising information at 41.6% and drafting or editing text at 40.9%. These are functions that have become embedded across professional and administrative roles in Canadian organisations, from financial services firms to public-sector agencies under the federal Canada Labour Code.

Sentiment is far from uniform, however. 30.9% of respondents report elevated concerns about how their organisation uses AI, and preferences for future roles pull in sharply different directions: 30.5% want a job with less AI involvement, while 20.4% want more. The ZipRecruiter report attributes much of that friction to employer-side governance failures, noting that the source of concern could be the employer not using AI, not having any guidelines or guardrails in place, or ethical and privacy concerns.

For Canadian HR leaders, that framing is useful precisely because it locates the problem within organisational control. Unclear policy, absent guardrails, and unresolved privacy questions are governance gaps, and governance gaps can be closed. The Canadian Human Rights Commission (CHRC) has begun flagging AI-related human rights considerations in the employment context, which gives People teams an additional compliance incentive to formalise AI use policies beyond retention alone.

The training gap follows the org chart

The split in training access is where the ZipRecruiter data becomes a workforce planning concern rather than a satisfaction metric. Senior leaders receive AI training at a rate of 83.5%. Below them, access drops to roughly half: 46.4% for individual contributors and 45.3% for entry-level staff. That is not a marginal difference in programme coverage; it is a structural imbalance in how organisations are equipping different layers of their workforce for a shift that affects all of them.

The Conference Board of Canada's workforce research has consistently found that training investment in Canada skews toward higher earners and more senior employees, a pattern this US data reinforces at the specific level of AI capability-building. The gap also mirrors the enjoyment data: employees with the most structured support report the biggest gains in how AI affects their experience of work, while those doing much of the hands-on operational work are left to figure it out independently.

The ZipRecruiter report notes that the lack, or insufficiency, of formal training in the workplace is leading many workers to teach themselves how to use AI tools, decentralising learning and putting the burden directly on employees. For HR, this creates a secondary problem beyond retention. When learning happens informally and without organisational oversight, visibility over which tools are being used, on which data sets, and to what standard is effectively lost. In sectors regulated under the Canada Labour Code, such as banking, telecommunications and interprovincial transport, that lack of visibility carries data governance and compliance implications that sit well above the People function.

Where the entry-level career pipeline is narrowing

The tasks AI handles most frequently are the ones that historically formed the first rung of professional development. Research, summarisation and first-draft writing were how many junior employees learned the substance of a business, built institutional knowledge, and developed professional judgement. The ZipRecruiter report states the consequence directly: many of the tasks AI is automating, like taking meeting minutes, summarising research, or drafting text, were all tasks that entry-level workers once owned, and there are now fewer opportunities to learn how to work, with the bar for entry-level workers rising.

For hiring managers at Canadian organisations, this changes the functional brief for junior roles. If AI absorbs the tasks that once trained new starters, the expectation on day one shifts toward candidates who can already evaluate, edit and verify AI output rather than simply produce a first version. Without deliberate role redesign, organisations risk asking more of entry-level hires while providing fewer structured development pathways, a tension that CPHR Canada and the Human Resources Professionals Association (HRPA) of Ontario have both identified in recent professional guidance on workforce readiness and skills-based hiring.

There is a pipeline dimension as well. Today's entry-level employees are the specialists and managers of the next decade. Reducing structured learning at the junior level now creates a capability deficit that becomes visible years later, when remediation is considerably more expensive. For organisations navigating succession planning in sectors facing demographic pressure, including healthcare, financial services and the federal public service, that long-cycle risk is not abstract.

Why the retention risk is deferred, not absent

A stated intention to leave is not a resignation, and the survey cannot establish how many of the 25.1% will act on it. The ZipRecruiter report is explicit that timing matters as much as intent, observing that in a job market defined by workers holding tightly to their current positions, this may not be a major issue for now, but when market activity heats back up, workers are positioned to act on their next move.

That deferred structure is the scenario worth modelling against. Statistics Canada's Labour Force Survey data for 2026 shows that voluntary quit rates in Canada remain below the levels recorded during the 2021-22 labour market peak. Low voluntary turnover in a cautious hiring environment can mask dissatisfaction that converts quickly into exits once more opportunities open up. Because the ZipRecruiter data covers US workers specifically, Canadian HR leaders should treat the directional signal seriously while validating it against their own stay interview, exit interview, and employee engagement data before drawing firm conclusions about domestic prevalence.

Employment and Social Development Canada (ESDC) has been tracking AI's effect on labour market transitions through its Future Skills Centre work, and that corpus provides a useful counterpart to US survey data for organisations wanting to ground workforce planning in Canadian evidence.

What HR leaders can act on this quarter

ZipRecruiter frames the current moment as an opening for employers to formalise and extend AI training across the full workforce. Several practical steps follow directly from the data, each of which Canadian HR teams can begin tracking this quarter without waiting for the labour market to shift:

• Audit training access by seniority level. Map who has received AI training, how much, and on which tools. If your internal distribution resembles the 83.5% versus 45.3% split in the ZipRecruiter data, you have a measurable equity and retention exposure that can be quantified and reported to the executive team.
• Publish clear AI-use guidelines. Concern about AI frequently traces back to the absence of guardrails and unresolved privacy questions. A written policy on approved tools, data handling standards, and accountability lines addresses that gap and provides a foundation for compliance with both provincial privacy legislation and the Canadian Human Rights Commission's (CHRC) developing guidance on AI in employment contexts.
• Redesign junior roles around judgement, not production. If AI drafts and summarises, give entry-level staff structured ownership over reviewing, verifying and improving that output, supported by regular feedback loops with more experienced colleagues. HRPA Ontario's competency frameworks for early-career HR professionals offer a model for how deliberate skill scaffolding can be built into role design.
• Recognise and credit self-taught AI skills. Many employees across Canadian workplaces are already developing AI capability independently. Capturing that learning in development reviews and skills inventories turns an informal burden into visible professional progression, and surfaces capabilities that would otherwise remain invisible to the organisation.
• Add AI adoption to stay and exit interview protocols. Tracking AI-related dissatisfaction as a discrete exit reason gives People teams an early indicator rather than a lagging one. Canadian HR Reporter, the principal trade publication for the domestic HR profession, has documented growing employer interest in AI-specific exit interview questions as a retention intelligence tool.
• Be explicit about entry-level AI fluency expectations. If junior roles now require the ability to evaluate and improve AI-generated output, say so in job descriptions and build onboarding support accordingly. Leaving that expectation implicit while raising the effective hiring bar creates a candidate experience problem and a diversity risk, given that AI fluency correlates with access to prior training, which is itself unevenly distributed.

Closing the training gap before the labour market turns

The ZipRecruiter data describes a workforce that is broadly open to AI but attentive to how employers introduce and govern it. The real divide is less between enthusiasts and sceptics than between employees who receive structured support and those left to develop capability on their own time and at their own cost. For Canadian organisations operating under the framework of the Canada Labour Code (for federally regulated sectors) or provincial employment standards regimes in Ontario, British Columbia, Alberta and Quebec, the absence of formal AI training provision is an emerging workforce risk with both retention and governance dimensions.

For HR leaders, the useful takeaway is that this risk is concentrated and measurable. Training access by seniority level, AI-related sentiment in engagement survey data, and the functional design of entry-level roles are all variables that can be tracked and reported against this quarter. Organisations that address the training gap while voluntary turnover remains low, as it currently does across much of the Canadian labour market according to Statistics Canada, are better positioned to retain capability when hiring conditions improve and workers begin to act on deferred intentions.