Australia’s AI future is here — but not for every business

Open data shows small and medium businesses are divided not just by AI adoption, but by skills, funding, trust and responsible use.

Soham Shrimant Rasam - s4176152

Australia’s AI future is arriving unevenly

Artificial intelligence is often described as a single wave that will reshape work, productivity and business. But the data suggests a less even story.

Some Australian small and medium enterprises are already using AI. Others are experimenting, waiting, unsure where to invest, or not yet aware of the tools available. The divide is not only about whether a business has access to AI. It is also about confidence, skills, funding, trust and responsible use.

This five-chart story uses open data from the National AI Centre’s AI Adoption Tracker to show where Australia’s SME AI divide is forming.

Main argument: Australia’s AI future may not be delayed by a lack of hype. It may be delayed by uneven readiness.

1. Adoption is visible, but hesitation remains large

The first divide is basic: some SMEs are already using AI, but many are still outside the adoption curve. In the most recent three-month view, broad use and limited use are higher than the all-time picture, but the largest group still says they do not intend to implement AI.

Source: National AI Centre, AI Adoption Tracker dashboard. Values manually extracted from the SME AI adoption view.

2. AI is not one tool

The AI divide also depends on what kind of AI a business is considering. Generative AI assistants have the highest current use in this dataset, but many other AI applications are still characterised by awareness without adoption, or limited awareness.

Source: National AI Centre, AI Adoption Tracker dashboard. Values manually extracted from the AI applications table view.

3. The divide is also about readiness

The data suggests that adoption is not only a technical question. Businesses report uncertainty about when to invest, lack of skills, lack of funding and concern about trusting AI decisions without ethical or bias issues.

Note: For speed, skills and funding, concern is measured as somewhat agree + strongly agree. For trust, concern is measured as somewhat disagree + strongly disagree.

4. Uncertainty is spread across several questions

The all-time response distribution shows why AI readiness is complicated. Neutral responses are high, but there are also substantial levels of agreement with skills, funding and speed barriers. Trust in AI decision-making is weaker, with more businesses disagreeing than agreeing that AI can make decisions without significant ethical or bias issues.

Source: National AI Centre, AI Adoption Tracker dashboard. All-time attitude responses.

5. Responsible AI practices are still catching up

The final divide is about responsible use. Even among businesses using AI, safeguards are not universal. Checking results before they affect customers is the most common current practice, but transparency and customer redress are weaker.

Source: National AI Centre, AI Adoption Tracker dashboard. Responsible practices are reported for businesses using AI only.

What this means

The data does not show one simple AI story. It shows several divides at once.

There is an adoption divide between businesses already using AI and those still outside the curve. There is an application divide between familiar generative tools and more specialised AI systems. There is a readiness divide around speed, skills, funding and trust. Finally, there is a responsible-use divide, where some safeguards are in place but others remain limited.

For policymakers, educators and business support organisations, this suggests that AI adoption should not be treated only as a technology problem. The question is not simply whether Australian businesses have heard of AI. The harder question is whether they have the skills, confidence, money and safeguards to use it well.

Data note

The charts use values manually extracted from the National AI Centre’s AI Adoption Tracker dashboard. Percentages represent survey responses from Australian small and medium enterprises. Some charts use derived measures, such as combining “somewhat agree” and “strongly agree” to represent agreement.

Acknowledgements

I acknowledge the use of ChatGPT to support initial brainstorming and topic selection for this assignment. No generative AI output was used as a final source of evidence. The data collection, analysis, visualisation design, R coding, interpretation, written discussion, referencing and final editing were completed independently by me.

References

Australian Government Department of Industry, Science and Resources. (2025, December 2). National AI Plan. https://www.industry.gov.au/publications/national-ai-plan

Baglin, J. (2026). Adding interactivity. In Data visualisation: From theory to practice. https://data-visualisation.stem.melbourne/adding-interactivity

Baglin, J. (2026). Data visualisation: From theory to practice. https://data-visualisation.stem.melbourne/

Baglin, J. (2026). Multivariate strategies. In Data visualisation: From theory to practice. https://data-visualisation.stem.melbourne/multivariate-strategies

Baglin, J. (2026). Storytelling with data. In Data visualisation: From theory to practice. https://data-visualisation.stem.melbourne/storytelling-with-data

National AI Centre. (2026). AI adoption. Australian Government. https://www.ai.gov.au/news-and-insights/ai-adoption-tracker/ai-adoption

National AI Centre. (2026). AI adoption tracker. Australian Government. https://www.ai.gov.au/news-and-insights/reports/ai-adoption-tracker

National AI Centre. (2026). AI applications. Australian Government. https://www.ai.gov.au/news-and-insights/ai-adoption-tracker/ai-applications

National AI Centre. (2026). How we collect data on AI adoption. Australian Government. https://www.ai.gov.au/news-and-insights/ai-adoption-tracker/how-we-collect-data-ai-adoption

OpenAI. (2026). ChatGPT [Large language model]. https://chatgpt.com/

R Core Team. (2025). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/

Sievert, C. (2020). Interactive web-based data visualization with R, plotly, and shiny. Chapman and Hall/CRC. https://plotly-r.com/

The Conversation. (n.d.). The Conversation style guide: Brand colours.

Wickham, H., Çetinkaya-Rundel, M., & Grolemund, G. (2023). R for data science: Import, tidy, transform, visualize, and model data (2nd ed.). O’Reilly Media. https://r4ds.hadley.nz/