Almost half (46%) of frontline workers surveyed by Humanforce have received no employer training, briefing or written rules on artificial intelligence (AI), despite 65% having used AI or an AI-enabled tool at work in the past year.
Among respondents who had used tools such as ChatGPT, Gemini, Claude or Copilot for work, almost one in four (24%) reported that their employer did not permit the practice, or that they did not know whether it was allowed. In healthcare, that figure rose to more than one in three (36%), indicating that AI is being adopted in high-stakes frontline environments before sector-wide guidance has been established.
Only 13% of workers reported receiving comprehensive training, according to the survey of 501 Australian workers across healthcare, hospitality and tourism, and retail and wholesale trade.
Humanforce CEO Clayton Pyne noted that while Australia’s AI debate still centres on desk-based workers, frontline employees are already using the technology on the job without adequate guidance. He stated that workers caring for patients, serving customers or moving between shifts require clear parameters on what information they can share with general-purpose AI tools, what outputs they can rely on and when responses require verification. According to Pyne, organisations that establish this clarity early will reduce risk and realise greater benefits from AI, including reduced paperwork, streamlined shift management and increased time for customers and patients.
Paperwork and administration ranked first across every sector as the area workers most wanted technology or AI to simplify (43% overall), followed by rostering and shift management (31%). Nearly half (49%) estimated they spend at least two hours in a typical week on tasks such as completing forms, managing timesheets, arranging shift changes, locating information and resolving pay or leave queries, while more than one in four (28%) spend four hours or more.
Pyne added that while frontline staff experience this administrative burden most directly, it also represents a compliance, cost and productivity issue for employers, making back-office operations a key area where AI can provide substantial support.
Pay accuracy remained a persistent issue. Two in five workers (41%) reported that their pay had been incorrect at least once in the past 12 months, and more than one in four (27%) stated it had occurred two or more times, an outcome that directly affects household budgets for employees paid by the shift.
The findings highlight a growing gap between frontline AI adoption and organisational governance across Australian workplaces. While employees are already turning to AI tools in high-stakes environments and seeking relief from persistent administrative and payroll pressures, the absence of clear policies and training leaves both workers and employers exposed to unnecessary risk. Bridging this divide will require organisations to establish practical guidelines and targeted back-office solutions that support frontline staff safely and effectively.
