SMEs have been promised they won’t be left behind as Australia adopts a new National AI Plan to supercharge productivity and transform the economy. 

Under the long-awaited federal government roadmap, businesses will rely on existing laws to govern AI use after a proposal for mandatory guardrails was abandoned. The government says the approach will balance the opportunities and risks posed by the rapidly-evolving technology.

“The National AI Plan is about making sure technology serves Australians, not the other way around,” said Minister for Industry and Innovation and Science Tim Ayres. “AI will help close gaps in essential services, improve education and employment outcomes and create well paid jobs in future industries.

Instead of introducing dedicated AI legislation, the government will set up a new AI Safety Institute to monitor for problems and advise industries.

Tech Council of Australia CEO Damian Kassabgi said the plan would support safe use without slowing innovation. “We believe this plan is a positive step towards recognising the opportunities that AI represents for our country, including 200,000 new jobs and $115 billion a year to the Australian economy by 2030.” 

How Does Small Business Fit Into This Plan?

Describing SMEs as the ‘backbone of Australia’s economy,’ the plan singled out small and medium-sized businesses for extra attention.

To ensure they remained competitive and efficient, the plan promised:

• Access to guidance and resources – government-backed tools to explore and implement AI technologies, via the $17 million AI Adopt Program
• Workforce training and upskilling – help for SME staff to develop AI literacy skills and foster innovation
• Supportive infrastructure – to reduce barriers to data centre development, providing business opportunities during construction and faster cloud access once complete
• Practical pathways to adoption – step-by-step examples of how to integrate AI into operations such as HR, marketing or customer service.

The plan stated more than one-third of SMEs had already adopted AI, and that Australia ranked third globally for consumer use of  popular AI tool Claude. But it noted a clear city-country divide, with only 29 percent of regional organisations adopting AI compared to 40 percent in metropolitan areas. It said closing this gap was a critical priority.

Employment Hero CEO Ben Thompson said supporting small and medium-sized enterprises in AI adoption was vital to driving productivity gains across the economy, in the midst of what he called a productivity crisis. 

“By prioritising regulatory gap analyses before new rules, expanding flexible data access, supporting SME training and digital capability, and avoiding unnecessary compliance burdens, policymakers can create the conditions for local businesses to thrive,” he said.

Getting Big Results For Small Businesses With AI

Thompson said AI held ‘enormous promise for the workplace,’ as one of the few economic levers the nation could pull to boost productivity without hiring more staff. He welcomed the commitment to data centre infrastructure and skills training, and said, with the right platforms, AI could act like an extra team member.

He described how the technology could widen access to expertise in areas like marketing, web design and data analysis, as well as pressure-testing business strategies and pricing, evaluating business plans and flagging cashflow risks before they became crises. “We’re already seeing this with hundreds of thousands of SMEs using AI-powered platforms like Employment Hero to access greater HR and employment expertise,” Thompson explained. 

He also noted AI was a vital tool for small and medium-sized businesses in navigating rising employment costs and stifling red tape. “Changes such as Single Touch Payroll and the new payday super laws add significant cost and operational burden, particularly for SMEs, who are disproportionately impacted,” he said. “If we can use AI to streamline and automate employment administration at scale, we believe it can significantly reduce labour administration costs. A 2 percent net reduction in these costs, which we believe is achievable, would equate to around $12.5 billion in efficiency gains for the sector.”

How Businesses Can Develop an ‘AI Police Plan’

In her role as an advocate for the responsible governance of AI, the Director of Ethical Advisory at The Ethics Centre, Aubrey Blanche, said she was heartened by the plan. But she believed an important step between policy and practice was missing.

“Some of the guidance assumes a level of bureaucratic infrastructure that I’ve never seen an SME have,” she explained. She said many small businesses wanted to introduce an AI safety plan but had limited background knowledge and didn’t know where to start or how to assess risk. “We need to invest in creating more resources for them to upskill so they’re actually capable of implementing the best-practice guidance in a context that makes sense.”

She said the frenetic pace of change meant owners should not wait until the new AI Safety Institute was established next year to develop their own AI protocols. In the interim, they could:

• Ask hard questions up front — before buying an AI tool, Blanche suggested pressing vendors on safety, ethics and data-handling procedures. She also said businesses should ensure products promising productivity gains were a good fit, after a study found 95 percent of AI pilot programs had failed to deliver valuable gains.
• Test it before you roll it out — businesses should not adopt an AI tool company-wide before confirming it works on a smaller scale, Blanche said. As an example, she recommended trialling HR software on case studies to check for biased or inconsistent outputs, especially in higher-risk areas like job application screening, before letting it loose across the whole business.
• Ensure staff know how to use it responsibly — Blanche said even the safest tool failed if misused, so staff had to be trained to recognise confidential information. “If you know half your employees are using Claude or ChatGPT, do you have an enterprise licence?” she asked. “Are you actually making sure what people are putting into the model isn’t being put in training data? I think those are issues that sound obvious to folks who work in this space every day but I actually don’t think they’re obvious at all because if you don’t know how the technology works, you wouldn’t even know to be worried about it.”

Blanche hoped the support the government promised would extend to AI governance and ethics training. “Also, I hope the SME community itself starts to organise and share best practices. I think all of that leads to the successful adoption of the National AI Plan.”