As artificial intelligence becomes embedded in Australian workplaces, the question is no longer if automation will affect jobs but who gets hit first, and how hard. Drawing on the latest data from the OECD, the World Economic Forum, and PwC, this article maps the automation risk landscape across occupations, revealing a picture that is more unequal and more urgent than most Australians realise.
Not all jobs face the same threat from AI. According to a 2021 OECD study of 98 skills and abilities, occupations involving manual, physical, and routine tasks score highest on automatability. Construction workers, farmers, and production workers sit at the top of the risk ladder, while roles requiring human judgement, empathy, and complex reasoning sit at the bottom.
On average across OECD countries, 27% of all jobs fall into the high-risk category (OECD, 2023).
Key insight: Positions that require repetitive physical activities are at the greatest risk of being automated. Positions that depend on human insight, compassion, and interaction are significantly less prone to automation. Instead of taking over these roles, AI is more likely to support these professionals.
By 2030, the WEF predicts 170 million new jobs will be created globally while 92 million will disappear. Technology roles like Big Data Specialists and AI Engineers are growing fastest, while clerical and administrative roles face the steepest declines.
Key insight: The fastest declining jobs are almost all clerical and administrative which are roles that involving processing information or following fixed procedures. These are exactly the tasks AI does best.
It is not just about which jobs disappear. The very nature of work is shifting. Today, nearly half of all work tasks are performed mainly by humans. By 2030, that share is expected to drop significantly, with technology taking on more and more of our daily tasks even within jobs that are not disappearing.
Key insight: By 2030, the share of tasks performed mainly by humans is set to fall from 47% to 33%. This means every worker’s daily role is being reshaped regardless of whether their job title changes.
Global data tells one story, but Australia has its own picture. According to PwC’s 2025 AI Jobs Barometer, based on nearly one billion job advertisements, Australian jobs are actually growing in AI-exposed sectors. Both roles where AI replaces tasks (automatable) and roles where AI enhances human work (augmentable) have seen strong growth since 2019.
Key insight: Contrary to fears, AI-exposed jobs in Australia are growing across every industry. Augmentable roles where AI assists rather than replaces are growing slightly faster, suggesting the near-term story is more about transformation than elimination.
Despite the growth story, many Australian workers remain uncertain or anxious about AI. Only 11% of Australian workers strongly believe AI will positively impact their roles which lower than the Asia-Pacific average of 16% and the global average of 17% (ADP Research, 2025).
Key insight: Only 11% of Australian workers strongly believe AI will benefit their role well below the global average. Meanwhile, 9% fear job replacement or feel uncertain about AI’s effect on their future. Building worker confidence through training and communication is as important as managing the technology itself.
ADP Research. (2025). People at Work 2025: A Global Workforce View. ADP. https://au.adp.com/resources/articles-and-insights/articles/t/the-impact-of-ai-on-job-responsibilities-in-australia.aspx
OECD. (2023). OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market. OECD Publishing. https://doi.org/10.1787/08785bba-en
PwC. (2025). 2025 AI Jobs Barometer: Australia. PwC Australia. https://www.pwc.com.au/services/artificial-intelligence/ai-jobs-barometer.html
World Economic Forum. (2025). The Future of Jobs Report 2025. World Economic Forum. https://www.weforum.org/publications/the-future-of-jobs-report-2025/
OpenAI. (2025). ChatGPT (GPT-4) [Large language model]. https://openai.com/chatgpt
This work was assisted by ChatGPT (OpenAI, 2025), which was used to help interpret assignment requirements and rubric criteria, explain concepts in simpler terms, assist with debugging and troubleshooting code, and clarify unclear instructions. ChatGPT was used as a learning and support tool rather than a source of original analysis. All data was obtained from the primary sources listed in the references, while the story, visualisation design, analysis, and conclusions were developed and decided by the author.