But decisive investments in the technology could deliver an additional $235 billion within a decade, a study has found, and would help the nation catch up to the rest of the world.
The Australian Academy of Technological Sciences and Engineering and consulting firm Kearney released the warnings on Friday from research that modelled three AI investment and policy scenarios.
The findings come as the industry waits for the release of the federal government's national AI plan, due before the end of 2025, and after federal Treasurer Jim Chalmers announced AI would become a national priority.
The study, called Unleashing Growth: Australia's AI Investment Blueprint, analysed investments in artificial intelligence technology from around the world.
It found other nations were on track to surpass $3 trillion in spending on AI in 2026, with a national average of $13.8 billion over five years.
By comparison, Australia had invested just more than $300 million in that time - significantly less than its nearest competitors Canada ($1.7 billion), the UK ($2 billion), Germany ($3.4 billion) and Singapore ($4.4 billion).
AI could boost productivity and efficiency across a range of fields, from farming to weather forecasts, academy president Katherine Woodthorpe said, but Australia would miss benefits if it did not increase its investment.
"Other countries aren't waiting and Australia risks getting left behind if we don't act decisively," Dr Woodthorpe said.
"AI is the tool that Australian businesses can turn into real productivity gains over the next 10 years but only if we manage to train enough people, build the data centres, and create the ecosystem of technology standards and regulations which are critical."
The study modelled three scenarios in Australia and found limited government action would deliver an $85 million boost to gross domestic product by 2034, while current levels of investment would boost it by up to $190 million.
A government investment of $5 billion over five years and greater support could unlock an economic boost of up to $235 billion, the report found, representing eight per cent growth.
The investment would be needed to upgrade Australia's supercomputing facilities, UNSW AI Institute director Sue Keay said, and to fund training for the next generation of technology workers.
"Often we hear a proposition put forward that Australia is too small, that Australia couldn't possibly invest enough money to be able to take advantage of the talent and the technologies that we are currently developing here," she said.
"This paper and the supporting economic modelling disputes the suggestion that Australia is helpless in this field and that we have to rely on purchasing AI off the shelf to help our productivity."
Developing homegrown AI models would also address brain drain in the sector, Dr Keay said, and provide investment confidence.
The report also recommended national AI training, purpose-built computing facilities, standards and regulations, and regional AI centres.
