It’s hard to ignore the buzz around artificial intelligence, which doesn’t seem to be winding down anytime soon
With 84% of Australian workers now using at least one type of AI on the job, it’s really easy for businesses to pat themselves on the back and assume they’re winning the tech race.
But here’s the thing. Just tracking how many people use these tools is overlooking a massive corporate shortcoming. We're celebrating rapid rollouts, but the leap from playing around with basic features to actually using AI safely and strategically is a bridge we haven't crossed yet. Instead, we need to move past the excitement of mere adoption and get serious about building real, structured capability and critical judgement in our teams.
The illusion of capability
It's tempting to think we're fully capable of AI just because everyone is experimenting. Sure, 84% of us are using AI at work, but the wide adoption doesn't mean we actually know what we're doing.
In reality, RMIT Online’s research has found that more than half of the workforce (54%) are still absolute beginners. Only 7% have reached an advanced level of AI literacy. The real conundrum is where our skills sit. Workers are twice as likely to be advanced in technical skills as in judgement skills. Meaning, lots of people can write a prompt, but very few can spot a hallucination, question a biased output, or realise when an AI-generated decision is creating legal or reputational risks.
For HR, this means adoption metrics are giving us a false sense of security. True capability means moving past just knowing how to use a tool and actually learning how to critically evaluate what it produces. I like to think about it as treating AI as an intern who’s just started in a role, leaving the judgement to us as ‘seniors’ reviewing their output.
The generational confidence misalignment
This literacy gap gets even trickier when you look at how different generations view their own skills. Younger workers naturally score higher on technical AI abilities than gen X and baby boomers, but they also face a pretty distinct risk: overconfidence.
Among millennials, 17% overestimate their AI literacy, and for gen Z, it's 21%. This overconfidence is risky because it makes it more likely they'll deploy AI without enough judgement, completely missing hallucinations or ethical missteps. On the flip side, 76% of baby boomers possess only beginner AI literacy, compared to 43% of millennials. They tend to be much more hesitant to engage with the tools at all.
So, HR professionals have a dual challenge on their hands. They need to temper the overconfidence of younger digital natives with ethical training, while simultaneously holding the hands of older, more hesitant cohorts to help them build foundational tech fluency.
The true cost of inaction – and the multi-billion-dollar opportunity
Ignoring this competency gap comes with severe financial and operational risks. Misusing AI brings very real costs in the form of rework, bad decision-making, and misplaced trust. In fact, one separate study found that 95% of respondents experienced an AI-related incident over the past two years, with the average financial loss sitting at a staggering $800,000. Despite these incredibly high stakes, only 48% of workers receive any kind of AI training from their employer, while a mere 11% get structured, ongoing support.
But the upside of getting this right is huge. If we could help just half of all our beginner-level workers improve to intermediate levels, we’d unlock a total economic dividend of $18.9 billion. To get there, HR needs to pivot toward providing structured, scenario-based training that lays out clear, practical rules for safe AI use at work.
At the end of the day, handing employees a shiny new AI tool is only step one. The real competitive advantage goes to the organisations that cultivate a workforce capable of critically evaluating AI outputs, navigating those tricky ethical grey areas, and applying the tech strategically to hit core business goals. It’s time for HR leaders to look beyond the basic adoption metrics and commit to robust, cross-generational upskilling. By putting just as much value on good judgement as we do on technical know-how, we can dodge those costly risks and finally unlock the true productivity potential of this AI revolution.
Rima Das is the director (workforce development) at RMIT Online
