Key insights
- 48 per cent of Australian employees who use AI admit doing it in ways that breach company policy.
- 68 per cent of Australian workers use AI at work and 35 per cent have had formal training from their employer.
- Only 30 per cent of Australians believe AI’s benefits outweigh the risks, the lowest result of the 47 countries surveyed.
Almost half of the Australian employees who use AI at work admit doing it in ways that are against their employer’s rules. More than half have relied on what it produced without checking whether it was right.
Those findings come from a KPMG and University of Melbourne study across 47 countries, and they describe the workforce that Australia’s National AI Plan has landed on. The same study found only 30 per cent of Australians believe AI’s benefits outweigh the risks, the lowest result of the 47. Whatever your rollout plan says, that is the room you are walking into.
Underneath the hidden usage sits something no policy document can reach. People are worried about being made obsolete, and worried that admitting a skills gap will speed it up. Workforce research now has a name for it: FOBO, the fear of becoming obsolete.
What is Australia’s National AI Plan?
Australia’s National AI Plan is the federal government’s roadmap for an AI-enabled economy, released in December 2025. It sets three goals: capture the opportunity through infrastructure, local capability and investment; spread the benefits through adoption, worker training and better public services; and keep Australians safe through regulation, responsible practice and international cooperation.
The plan has moved past its launch announcement. The government committed $29.9 million to establish an AI Safety Institute by early 2026, to test emerging AI risks and support regulators, and the Future Skills Organisation is building digital and AI competency units into the training system. The plan names workers and unions as part of the transition, with an expectation that adoption is transparent, safe and responsibly managed.
What does the National AI Plan mean for workplaces?
Spreading the benefits is the goal with your name on it. Workforce training sits inside the plan’s core objectives, which gives HR leaders cover to take the AI conversation past productivity targets and into capability, fairness and job quality.
The gap is measurable. EY’s Australian AI Workforce Blueprint, a survey of 1,003 Australian workers, found 68 per cent already use AI at work while 35 per cent have received formal training from their employer. Seventy-two per cent worry about breaching data or regulatory rules when they use it. The experimenting carries on regardless, unsupervised and unrecorded.
What is FOBO and why does it stall an AI rollout?
FOBO is the fear of becoming obsolete: the worry that your skills, role or professional standing will be made irrelevant by technology. In a rollout it presents as resistance to the tool, when the driver is self-protection. Someone who suspects a system might replace them will not volunteer to be trained on it in front of their manager.
-
30%
Australian workers worried AI will replace their job
Finder’s Consumer Sentiment Tracker, 669 employed Australians surveyed in February 2026. The figure rises to 38 per cent among Gen Z.
-
58%
US workers who fear becoming obsolete through a skills gap
The Harris Poll surveyed 15,357 US adults for ETS in late 2025. No equivalent Australian measure has been published.
Finder’s Consumer Sentiment Tracker found 30 per cent of employed Australians are worried AI will replace their job, rising to 38 per cent among Gen Z, and a Harris Poll of 15,357 US adults for ETS found 58 per cent feared becoming obsolete through a lack of skills.
Australia comes to this less prepared than most. KPMG’s Australian findings show 24 per cent of Australians have undertaken any AI training or education, against 39 per cent globally, and more than 60 per cent report low knowledge of AI. Those are the conditions FOBO grows in.
-
Training sessions go unattended
Turning up admits a capability gap in front of colleagues, so people find a reason to be elsewhere.
-
Legacy tools stay in service
Teams keep the old process running long after a faster option is available to them.
-
The questions change shape
Staff ask whether the tool will take their job far more often than they ask what it can do for their work.
-
Decisions travel upward
Managers push AI choices to someone more senior rather than own a tool that might turn out to be the wrong pick.
What does a rollout without FOBO look like?
The same software, introduced two ways, produces two different teams a year later. Take a customer service team adopting a tool that drafts replies and summarises calls.
| Rollout decision | Team A | Team B |
|---|---|---|
| Opening message | A new average handling time target, with the tool attached to it | What the tool drafts, what it is not allowed to decide, and where customer data goes |
| Who tested it first | The vendor and the project team | Frontline staff, before any target was set |
| Link to performance ratings | Never stated, so staff assumed the worst | Ruled out in writing by the team’s manager |
| Twelve months on | Use has narrowed to the few staff who never objected | Still in use, with checking the output now a team habit |
Illustrative comparison, not a documented case study.
How do you defuse FOBO while implementing the plan?
Be specific about what is changing, what is protected, and where each person fits. Five commitments do most of the work, and each one can be checked at a leadership meeting.
| Commitment | What it looks like in practice |
|---|---|
| Name the fear | Say in the launch message that worry about obsolescence is reasonable, then say what the organisation is doing about it. |
| Map roles, not tools | Publish, for each role family, what AI will do, what it will not do, and how the role is expected to change over 6, 12 and 24 months. |
| Train towards better work | Position upskilling as a route to more valuable work, with the hours allocated inside the working week rather than added to it. |
| Guarantee human review | Give every AI-influenced employment decision a named human reviewer and a stated appeal path. |
| Update it quarterly | Revisit the role maps every quarter so people can see their place in a plan that is still moving. |
FOBO does not disappear because an employer adopts AI well. It settles into ordinary uncertainty when leaders keep saying what is changing and what is staying put. That is the cultural work behind the National AI Plan, and it is the part no policy document can do for you.
FAQs: the National AI Plan and FOBO at work
Does the National AI Plan create new rules for employers?
Not as a single AI law. The plan is a policy roadmap, and the government has signalled it will work through existing regulators and targeted reform rather than one new statute. Employers still carry their duties under privacy, employment and work health and safety law whenever they deploy AI, so building governance, consultation and training now is the practical move.
What is FOBO at work?
FOBO is the fear of becoming obsolete: the worry that your skills, role or professional standing will be made irrelevant by technology. It shows up as avoidance of AI training and attachment to legacy tools. The acronym has an older meaning in decision-making research, fear of a better option, which is unrelated to this use.
How should HR leaders prepare for AI in the workplace?
Start with an audit of where AI is already in use, including the tools staff adopted on their own. Name an accountable owner, publish a policy, and consult staff before AI changes how work is assessed, allocated or monitored. Training is the biggest measured gap in Australia: EY found 35 per cent of workers have had formal AI training from their employer.
Last verified August 2026. Implementation of the National AI Plan is ongoing and the funding, timing and remit of the AI Safety Institute may change. Check industry.gov.au before republishing those details.
