Welcome back to Neural Notes, a weekly column where I look at how AI is influencing Australia. In this edition, in a conversation with Lee Hickin, executive director of the National AI Centre, we discussed why the country’s advantage may lie less in building frontier models and more in applying AI to industries where it already has deep expertise and unique datasets. For the past two years, the global AI race has largely been a contest over models and compute power. Governments are investing in sovereign capability, tech giants are pouring billions into data centres, and startups are chasing the next breakthrough model. And there certainly has been no shortage of debate locally about how Australia can compete with larger overseas players in this environment. But Hickin’s view is this may not be where Australia’s edge sits. Related Article Block Placeholder Article ID: 331868 Neural Notes: The blue caviar problem AI was built to solve Tegan Jones “Where does Australia have data and insights and knowledge that is unique to us,” he said to SmartCompany during the launch of ARM Hub’s Propel-AIR 2.0 robotics and AI accelerator in Brisbane. Smarter business news. Straight to your inbox. For startup founders, small businesses and leaders. Build sharper instincts and better strategy by learning from Australia’s smartest business minds. Sign up for free. * indicates required Email Address * By continuing, you agree to our Terms & Conditions and Privacy Policy. Rather than trying to replicate the likes of OpenAI, Anthropic or Google, Hickin points to an opportunity in applying AI to industries where Australia already has a wealth of knowledge and large volumes of data remain underutilised. Where Australia already has an AI edge “I clearly see agriculture,” Hickin said.  “I see biochemistry and the medical industry, and of course, resources and mining.” In other words, the same sectors that have largely underpinned the Australian economy for decades. Agriculture generates vast amounts of data across crop yields, soil conditions, weather patterns and satellite imagery. Mining and resources companies hold decades of geological surveys, exploration records and operational data. Medical research institutions manage extensive clinical datasets tied to Australia’s health system. “The question isn’t just how to make that industry do more… but how to develop insights that are viable to a global market,” Hikin said. Related Article Block Placeholder Article ID: 332633 Neural Notes: Australian AI startups may face the same defence dilemma as OpenAI, Anthropic Tegan Jones There’s an argument in the idea that as foundation models become more widely available, more of the value is being created in the layers built on top of them. We’re already seeing this extensively with the myriad of ‘AI startups’ that have popped up to disrupt specific industries such as law, medicine, HR and more. Companies are combining general-purpose models with proprietary datasets, domain expertise and specific workflows to create products that are harder to replicate. “The opportunity space is where there is a real gap in the applied use of data and knowledge and information in a rapid way,” Hickin said. That dynamic plays to Australia’s strengths, particularly in industries where data has been accumulated over long periods and is tied to physical systems, environments or regulatory frameworks. “The opportunity… creates a value that Australia could contribute back to the market, because we are world leaders in some of those sectors,” Hickin said. Capturing the value If that is the case, the more important issue may be who ultimately captures the value created from those applications. In the past, Australia has at times struggled to commercialise its research and industry expertise fully. AI may not be an exception. If the underlying models are developed offshore, and the platforms through which they are deployed are controlled by large global companies, there is a risk that much of the value generated from applying AI to Australian industries is also captured elsewhere. At the same time, access to high-quality datasets is uneven. Large incumbents often hold the most valuable data, while startups and smaller players may struggle to access it in usable forms. That creates a potential issue between the theoretical opportunity and an actual ability to act on it. For now, Hickin’s view is that the opportunity is still very much there, but it will depend on how effectively businesses engage with it. “Give industry the opportunity,” he said. “Here’s the technology, here’s the data you understand, do something amazing with it.” Stay in the know Never miss a story: sign up to SmartCompany’s free daily newsletter and find our best stories on LinkedIn.