Nearly half of Australian workers are now using AI at least multiple times a week but only 13% think it’ll improve how they do their job. That gap, between how much AI has crept into daily work and how much people trust it, is the story in ADP Research’s new People at Work 2026 report.
Globally, half of workers now use AI multiple times a week, and one in five use it almost daily. In Australia, 44% of workers are using AI at least a few times a week, and 16% use it nearly every day.
Younger workers and knowledge workers are leading the uptake. Those aged 27 to 39 are the most frequent users. And those in knowledge roles feel more confident about AI’s impact on their job than people in skilled-task or repetitive-task roles.
But confidence hasn’t caught up with adoption. Only 13% of Australian workers expect AI to have a positive impact on their job responsibilities over the coming year.
Part of that gap may come down to who’s expected to make AI work. Many EAs and admin professionals report being informally handed responsibility for AI adoption in their organisation – becoming the person who champions new tools and trains others on them – often without any additional recognition or investment in developing those skills.
If you’re carrying that responsibility without support, it’s not surprising confidence lags behind use.
The productivity paradox
Daily AI users are four times more likely to feel less productive than people who don’t use AI at all, even though they report higher engagement and lower stress. That’s not a contradiction so much as a measurement problem.
As AI takes over routine, checklist-style tasks, workers shift toward longer-term, more strategic work – the kind that’s harder to point to at the end of a day and say “I did that.” The feeling of having achieved less can creep in even when the actual impact is greater.
There’s also a simpler, more exhausting version of this gap. Keeping up with what AI can now do takes real time, and that time is rarely built into anyone’s day. It’s easy to default to doing something the way you’ve always done it, not because AI can’t help, but because there’s no space to stop and learn a faster way.
For EAs, that tension will likely sound familiar. Much of the role already involves work that’s difficult to quantify – anticipating problems before they happen, protecting an executive’s time, making judgement calls that never show up on a to-do list.
Add AI handling more of the visible, checkable tasks, and the gap between what you’re actually delivering and what feels measurable only widens.
What frequent use does seem to change
It’s not all downside. Daily AI users report notably lower negative stress than non-users (11% compared to 23%) and stronger team dynamics, including feeling more part of a team and more confident their job is secure.
The engagement is real, even where the productivity feeling is complicated.
What this means day to day
The disconnect between adoption and confidence isn’t something an individual EA can solve alone, but it’s worth naming rather than quietly absorbing.
If you’re using AI daily and still ending the week uncertain whether it’s actually made you more effective, that’s not a personal failing. It’s a documented, widespread pattern, and one made harder if you’re also the one expected to figure the technology out for everyone else.
The shift toward less tangible, more strategic work is exactly the direction the EA role has been heading anyway. The measurement challenge that comes with it is now something the whole workforce is catching up to, not something unique to the job.