Ask most Australians what worries them about AI and they'll say losing their job. But the data tells a different story. The bigger danger isn't a robot showing up to your desk. It's your employer not preparing you for what comes next.
New research from Indeed, Randstad, and Salesforce paints an uncomfortable picture: Australia has one of the widest AI training divides of any country studied. The workers who need upskilling most are getting it least — and the gap is growing.
The training divide in numbers
A survey of Australian workers and executives found that 72 per cent of executives received AI-related training in the past year. For managers, it was 32 per cent. For frontline workers, just 17 per cent.
That four-to-one ratio between the C-suite and the shop floor is not just a training problem. It's an information asymmetry. Executives understand what AI can do because they've used it. Workers don't — and 56 per cent of Australian businesses have no plans to retrain them.
Randstad Australia identified what they call a 23-point "AI blind spot": 68 per cent of employers expect AI to affect a high proportion of work tasks, but only 45 per cent of workers agree. When your boss sees a wave coming and you don't, you're not going to brace for impact.
Indeed's Hiring Lab found the gap extends to hiring too. As of February 2026, 6.2 per cent of Australian job postings mention AI — double the rate from a year ago. But two-thirds of those postings come from just one per cent of employers. AI adoption is concentrating in a small number of large firms, leaving workers at smaller businesses even further behind.
The jobs most exposed and least prepared
So which Australian workers are caught in this gap — high AI exposure, low training access?
General clerks sit right in the crossfire. With an AI exposure score of 7.0 out of 10 and 286,600 people employed, it's one of the largest and most vulnerable occupation groups in the country. According to Jobs and Skills Australia, 71 per cent of clerical tasks have automation potential. The median weekly earnings sit at $1,438 — well below the national average. These aren't the workers getting sent to AI workshops.
Accounting clerks face a similar squeeze. AI exposure of 7.2, median age of 45, and flat employment growth of 1.3 per cent over five years. Their work — data entry, reconciliation, invoice processing — is exactly what current AI tools do well. Yet they're among the least likely to be offered structured training.
Bookkeepers carry an AI score of 7.1, a median age of 52, and median earnings of $1,400 per week. The occupation is 84 per cent female. This is the demographic triple that falls hardest into the training gap: older, lower-paid, and female-dominated roles are consistently the last to receive employer-funded AI training.
Jobs and Skills Australia's Generative AI Capacity Study found that 55 per cent of workplace tasks could be performed by people using AI, while 15 per cent could be fully replaced. The key phrase is "by people using AI." The technology doesn't replace the work — it replaces the worker who can't use it.
Who is actually adapting
Not every occupation is stuck. Software and applications programmers (AI score: 6.7, 195,400 employed) are being forced to adapt because their industry demands it. Indeed found that 43 per cent of software development job postings now mention AI skills. If you're a developer who hasn't touched a large language model yet, employers have noticed.
Advertising and marketing professionals (AI score: 6.2, 102,500 employed) are in a similar boat — 17 per cent of their job postings reference AI. Content generation, campaign optimisation, and audience targeting are all areas where AI tools have become standard.
AI engineer is the number one fastest-growing occupation in Australia, according to LinkedIn's Jobs on the Rise 2026. Eight in 10 business leaders say they'd prefer to hire someone comfortable with AI tools over someone with more experience but no AI skills.
But here's the catch: even in these adapting sectors, the opportunity is unevenly spread. Indeed's data shows that top-quintile employers (by posting volume) mention AI in 6.7 per cent of job ads. Bottom-quintile employers manage just 3.1 per cent. If you work for a small business — and most Australians do — your employer is probably not pushing AI adoption at all.
A two-speed workforce
What's emerging is not a simple story of jobs being replaced. It's a story of two workforces splitting apart.
On one side: workers who get trained, who understand AI tools, who use them to work faster and take on higher-value tasks. JSA calls this "augmentation" — and it's happening in every industry. On the other side: workers whose roles are slowly hollowed out because they weren't given the chance to adapt.
As Griffith University's Professor David Tuffley told SBS News: workers who refuse or aren't given the opportunity to engage with AI risk becoming "less and less useful employees" over time.
This isn't a future scenario. It's happening now. Only 41 per cent of Australian workers say their workplace is prepared for AI, well below the global average of 48 per cent. And between 16 and 40 per cent of workers are entirely disengaged from AI — they don't use it and don't feel they need training. That disengaged group skews older, lower-income, and toward manual occupations.
Meanwhile, the occupations with the lowest AI exposure are under the least pressure to change at all. Electricians (AI score: 2.8, 195,900 employed) and other trades remain largely insulated — their work requires physical presence, hands-on problem-solving, and adapting to unpredictable environments that AI can't replicate. The irony is that the workers who are safest from AI are also the ones with the least incentive to train for it.
What this means for your career
The pattern across Australian occupations is clear:
High exposure, low training access — general clerks, accounting clerks, bookkeepers, payroll clerks. These workers face the highest risk not from AI itself, but from being left behind while their roles change around them.
High exposure, adapting — programmers, marketing professionals, ICT analysts. Their industries are pushing adoption and their job postings reflect it. They're upskilling because the market demands it.
Low exposure, low urgency — electricians, plumbers, concreters, carpenters. Physically demanding, unpredictable work that AI can't easily automate. Training pressure is minimal.
The uncomfortable truth is that 74 per cent of Australian workers say developing AI capabilities is important to them. The appetite for learning is there. What's missing is access — employers aren't investing, smaller businesses aren't adopting, and government programs are only just ramping up.
If you want to understand where your own occupation sits, check your AI exposure score or browse the full occupation rankings to see how your role compares to others across Australian industries.
The question isn't whether AI will take your job. It's whether you'll be ready when AI changes it — and whether anyone is helping you get there.