A national strategy for artificial intelligence (AI) will back three key industry sectors, but experts warn that Australia is still well behind the curve when it comes to building an industry that will require up to 161,000 new specialised workers by 2030.

Produced by the CSIRO’s Data61 data-science arm, the Artificial Intelligence Roadmap is filled with heady projections and aspirational forecasts including the estimate that AI could be worth $22.17 trillion to the global economy by 2030.

The technology will form part of a digital-technologies mix that will, the report predicts, be worth $315 billion to Australia’s economy by 2028 – but just how much of that pie Australia can grab will depend on its ability to turn AI ambitions into real-world deliverables.

The country’s efforts will, Data61 advises, be best spent targeting three key industries including health, aged care and disability services; towns, cities and infrastructure; and natural resource management.

The choice of these possible specialisations “are based on our existing strengths and capabilities, and opportunities to solve problems and export AI-driven solutions,” Minister for Industry, Science and Technology Karen Andrews said in launching the report at the inaugural Techtonic National AI summit in Canberra.

AI may be “a source of mystery and anxiety” for many Australians but ultimately “AI will mean more jobs, not less”, Andrews said, calling the technology “crucial” to the government’s ongoing commitment to creating 1.25 million new jobs by 2024.

“AI will support, enable and augment the human experience – not replace or destroy the things we value.”

Building a domestic capability

Although Australia has established strong capabilities in “core AI-related fields” within universities, research organisations and companies, the report notes, the country will need to develop an additional 32,000 to 161,000 AI specialist workers by 2030.

Meeting this target will require increased science, research and technology investment around high-level strategies including technological specialisation, mission-directed research, and business and knowledge ecosystems that span “jurisdictional boundaries”.

Development of AI will be “a global effort”, the report says, noting that “no one country will drive or decide how the forthcoming AI transformation happens.”

“The question is how we can best position our nation to adapt and capitalise on these changes for the benefit of all Australians.”

Buried amongst several aspirational use cases – highlighting the potential benefits of applications like AI-powered agricultural robots, autonomous mining vehicles and predictive maintenance of the Sydney Harbour Bridge – is a statistic that could weigh heavily on the nation’s ability to execute Data61’s AI vision.

A survey of 72 Australian AI start-ups, conducted during preparation of the report, suggested that 84 per cent are “adapting or developing novel algorithms” instead of simply consuming AI tools built by others.

That’s a key measure of innovation in an industry that will, given the multi billion-dollar AI investments and national strategies of countries like the United States, Germany, Canada and China, increasingly be comprised of leaders and followers.

“The success of our industries of the future will be determined by whether AI is simply used to cut costs, or whether we take full advantage of this powerful technology to grow new opportunities and create new value,” CSIRO CEO Dr Larry Marshall wrote in introducing the report.

“The global race to lead in AI is well and truly underway.”

An advantage already lost

Yet by some accounts, clear winners have already emerged.

“There are indications that the window for competitive differentiation with AI is rapidly closing,” Deloitte writes in its latest State of AI in the Enterprise executive survey – in which Australia came dead last in terms of the per centage of companies looking to use AI for competitive advantage.

Fully 50 per cent of Australian companies were just hoping to use AI to catch up and just 22 per cent would use AI to widen their lead or leapfrog ahead – compared with 21 per cent and 55 per cent, respectively, in China.

Two-thirds said AI will have already transformed their industry within 5 years, noting that “as AI technologies become easier to consume and get embedded in an increasing number of products and services, the early-mover advantage will rapidly diminish.”

That means the next decade will be crunch time for AI – and for Australia to have any say in its evolution, it will need to move quickly to execute on the AI Roadmap.

That could prove harder than many want to believe, warns distinguished professor Mary-Anne Williams, director of The Magic Lab within the UTS Centre of Artificial Intelligence.

“Australia lacks both significant IT manufacturing capability and distinctive AI software offerings,” she wrote in a recent journal article evaluating Australia’s AI opportunities.

“Australians like to believe that Australia punches above its weight in a wide range of areas, but there is little independent evidence to suggest that Australia punches above its weight in AI.”

With Australia “one of the worst performers in the OECD” on knowledge transfer, she writes, “the reality is that Australia is punching well below its weight in areas that determine a nation’s future safety, security, productivity and prosperity.”