How can Australia become an AI leader?

If 2023 was a year of ‘Artificial Intelligence (AI) awakening’ due to the rise of generative AI (GenAI), 2024 is a year of ‘AI practicality’ as organisations around the world, including in Australia, look to turn hype into reality.

According to the IDC Asia/Pacific AI Maturity Study 2024 commissioned by Intel, Australia is currently an AI innovator (stage 3), which means the country has made concerted efforts in planning and managing AI initiatives through well-laid technology infrastructure and data management strategies.

Compared to the Asia–Pacific (APAC) region’s AI maturity, Australia’s scores are above average for the enterprise and government dimensions, but only slightly above average for the socio-economic dimension. AI spending growth in Australia is forecast to have a compound annual growth rate (CAGR) of 25.4% by 2027, or around $12.67 billion. This CAGR is the second highest among the eight markets studied and is being driven by investments in software and infrastructure. With government initiatives and widespread AI adoption across various industries, Australia is poised to become a leader in AI adoption.

Yet many Australian organisations have realised that deploying AI is no easy task. The IDC study found that while a majority of mostly large Australian enterprises surveyed are investing in AI, only 7% believe their adoption of AI is core to their organisations’ competitiveness, highlighting the difficulty of successfully deploying AI in Australia.

Addressing Australian organisations’ AI bottlenecks

While AI is a complex technology with many facets to consider, here are a few key areas that I believe are top priorities for Australia.

To start on the right foot, organisations must begin with the business challenge or outcome they want to solve or achieve — not by chasing the latest buzzword. What may benefit now may not be a GenAI model, but be using interpretive or predictive AI to speed up business processes and arm employees with data-driven insights. Australian businesses must understand that AI is a technology with a wide spectrum of methods and tools to achieve different goals, not the goal itself.

This leads us to another top AI challenge — that being the perception that ‘the cost of ownership is too high to justify’, according to the IDC study. Since last year, there has been a rush towards obtaining more graphics processing units (GPUs) to power AI, resulting in a global shortage and high prices for these processors. However, contrary to the conventional belief that AI must use GPUs, the reality is that there is a variety of AI chips — be it GPUs, central processing units (CPUs), field programmable gate arrays (FPGAs) or application-specific integrated circuits (ASICs) for different AI use cases.

In fact, many common AI workloads can be done by CPUs efficiently without requiring additional GPUs, while the same CPUs can be used for other workloads like networking and storage at the same time — saving tremendous cost and energy consumption.

Another important trend to note is where data resides. By 2025, an estimated 75% of global data will be created and processed at the edge (eg, in factories, hospitals, etc), outside of a traditional data centre or cloud environment. This requires businesses to manage data and compute where it makes the most sense, be it in the cloud, a PC or at the edge, to manage the different requirements of model training and inference as AI evolves and expands.

Australian organisations need a technology foundation that is flexible and scalable — a heterogeneous compute environment and open ecosystem that allows them to choose the right hardware and software tools for the right jobs, regardless of vendors, will be key to lowering costs and achieving better reliability, performance and security for AI down the road.

On macro-economic challenges: government regulations and AI talent

While businesses are critical to driving AI adoption and innovation in Australia, the nation also relies on the macro environment to drive AI forward.

From a government perspective, the Australian Government has been proactive in developing AI policies and encouraging investments. As part of the 2023–24 May Budget, the government allocated $101.2 million to support Australian companies to integrate quantum and AI technologies into their operations, and has emphasised its commitment to fostering innovation and economic growth through AI. This investment is also in line with the government’s stated goal of creating 1.2 million tech-related jobs by 2030.

While the Australian Government takes a less restrictive, innovation-friendly approach to AI, incremental progress in establishing AI regulation could also lead to risks of misuse, unfocused AI technology investments and disparities in AI advancements.

On the other hand, while Australian universities have led in AI-related research and demonstrated strong expertise in AI, like many of its peers in the APAC region, the nation does not have enough technology talent.

What is more, people with relevant experience in AI are rare and expensive, which often puts them out of reach for small and medium-sized enterprises (SMEs) that make up most of Australia’s economy. Some enterprises also inevitably must seek technology talent offshore. Australia’s relatively small and geographically dispersed population also means that technology talent is unevenly distributed across the country as well as sectors.

The government and industries must therefore make nurturing and upskilling technology talent a priority. There is the need to not only train, but also bridge academic knowledge and practical AI skills. One area to also consider is to expand AI understanding in the non-technical workforce. With the rise of low-code/no-code AI tools, the non-technical workforce’s understanding and adoption of AI will be important in advancing Australia’s AI competitiveness as well.

Becoming an AI leader

With its many advantages, Australia is well-poised to join the fast AI bandwagon in the years ahead. SMBs and enterprises alike will benefit from cost savings and productivity gain, and the nation as a whole will be more competitive.

Along with this significant AI opportunity and growth, complexities will also rise. The types of AI will continue to evolve, and the approaches to AI will have to adapt and change. This means that the technology, government and societal infrastructures that we build today need to be flexible and scalable as well. As the next few years will be a defining period for countries to build out their AI foundation, how Australian businesses approach, invest and build AI, how the government refines the regulations and safeguards societal interest, as well as how the AI talent pool is nurtured and retained, will be pivotal for Australia to become a leader in AI.

Top image credit: iStock.com/pcess609