New research finds one in seven employees trust AI over colleagues, as ELMO's people chief warns guidance can't keep pace
One in seven Australian employees – roughly 1.5 million people – now turn to artificial intelligence (AI) for work advice before asking a manager or colleague, according to new research from ELMO Software's Employee Sentiment Index (ESI).
The finding, drawn from a survey of 1,068 Australian employees conducted by YouGov between January and June 2026, points to a shift in how workers seek guidance that HR leaders everywhere are watching closely, even where the underlying data is Australian.
A further 44% of employees – around 4.7 million people – say their choice between AI and a human depends on the task, while just 42% still default to a manager or colleague first. Among senior leaders and business owners, trust in AI nearly doubles: 26% say they would ask an AI tool before a person. Employees cite speed (57%), convenience (44%) and better-quality answers (44%) as their main reasons, well ahead of social factors such as not wanting to bother a colleague (28%).
A trust dividend, not a trust deficit
For Anne Tosky, global chief people officer at ELMO Software, the finding reflects a broader change in how people default to technology for advice.
"If I think about it – I'm going into the doctor, but I'm free ChatGPT-ing my symptoms first, which I'd be asking. It's my kind of annual checkup," she told HRD, describing a pattern she sees carrying over into the workplace. "Going to AI first is kind of your – okay, tell me what's going on here, what should I do? And that extra sounding board."
Tosky said the trust her own organisation has built with AI did not happen by accident. "With the right foundations in place, the trust is there and warranted," she said, pointing to ELMO's early investment in AI policy, guidance on when to use a human in the loop, and a "Change Champion network" of employees who help build AI capability at a grassroots level.
The company has also built an internal HR agent, named Poppy, that answers policy and payroll questions without straying into advice. "She runs across our HR policies and systems and gives you accurate advice … but doesn't step into advice," Tosky said. "Need to be thinking about that, because people are going to do that [use AI] on their ChatGPT anyway. So how to get the best advice in the right way to people."
Employer expectations are outrunning employer guidance
Trust in AI is rising faster than organisations' ability to manage it responsibly, according to ELMO Software President Joseph Lyons. "With adoption high, the question now is how organisations can systemise AI so it's a trustworthy single source of truth," Lyons said.
"Australian organisations are still struggling with unregulated shadow AI tools being used, and many employees aren't always comfortable with how AI is used in the workplace." The ESI data backs this up: while 66% say responsible AI use is valued at their organisation, only 57% say it is clear when AI should and shouldn't be used, and almost half (49%) have felt uncomfortable with how AI is used at work.
Lyons argued the answer is not caution but structure. "Organisations need to consider not only when employees should use AI, but how to validate its output and when it should be trusted over a human," he said. "Like any widespread change, this takes a consistent, planned approach. Regulated use won't happen organically."
Wellbeing hasn't caught up with adoption either
The ESI also found the shift toward AI is unfolding against a backdrop of rising strain: 44% of employees reported feeling burnt out in H1 2026, up from 39% in H2 2025, and confidence in personal economic security fell to 28% from 37% over the same period.
Dr Amantha Imber, organisational psychologist, AI consultant and ELMO partner, said the response to that pressure sits with employers even where the cause does not.
"While employers can't control the broader economic landscape, they can consider how they support their team's overall physical and mental wellbeing in these conditions, whether through flexible working arrangements, mental health days or extra support," she said.
For HR leaders navigating growing reliance on AI, the point is a reminder that a trusted AI tool is not a substitute for a workplace employees feel supported in.
The culture question HR can't outsource
Beyond governance, Tosky flagged a subtler risk: what happens to workplace relationships and early-career development when the default sounding board is a chatbot rather than a colleague.
"Without that kind of connection, storytelling, sharing, learning on the job, just asking the AI – you're not going to get that," she said, adding that preserving culture alongside AI transformation is now central to her role.
The pressure on HR to hold both priorities at once – speed of adoption and strength of culture – is intensifying globally. Trust, not technology, determines AI success. Governments are also starting to respond: Australia's federal government has moved to introduce national AI standards in a trailblazing move.
For HR leaders in every region, the ELMO data offers an early signal rather than a local curiosity: employees are not waiting for policy to catch up before they trust AI with decisions that once went to a manager. The organisations that get ahead of that shift, Lyons and Tosky both suggest, are the ones treating trust as something to be systemised and taught – not assumed.