Opinion
Australia needs a more intelligent approach to AI from both major parties at the next election, writes Bede Moore.
In his Q4 earnings call of 2022, Satya Nadella outlined the details of Microsoft’s ambitious Enterprise Metaverse strategy, a vision for transforming how businesses engage with digital environments. It was a bold and ambitious strategy. Today, that strategy lies in the dustbin. It has taken just two years for the company to replace Enterprise Metaverse with the announcement of $100 billion in capex, chiefly in global data centres, hardware for AI workloads and initiatives aimed at scaling generative AI.
It is a mark of Nadella’s strength of leadership that he can acknowledge such a large sunk cost. It is a leadership trait we must all now emulate – even talk of two years feels like eons in the context of the latest AI advances. Since the release of Claude’s 3.7 Sonnet a few weeks ago, software engineers and product managers are feeling distinctly wary about their future job prospects. Thoughtful English-language prompting is already enough to build you, say, a transacting website – a functioning business built by one person with some prompts.
You could be forgiven for thinking by this stage in 2025 that you’ve already witnessed enough disruption for the year. Nobody wants to read another word about the Trump Presidency, but its mere mention is enough to suggest epochal change.
Very few people predicted the tectonic shifts that would emanate from Washington in the early months of this year, but they have captivated the world’s attention and left political and commercial leaders scrambling to comprehend the new environment.
That shock might explain Australia’s flat-footed response to the latest developments in AI and their implications for the Australian business environment. At least this is what you would say if you were being generous.
A more reasonable analysis is that Australia’s political leadership is asleep at the wheel as we hurtle towards a fourth industrial revolution. The first result on the Liberal Party’s website for “Artificial Intelligence” yields a speech by Peter Dutton in Mount Waverley. The term “Artificial Intelligence,” appears just once – with zero actual substance – in their 44-page “Plan” for the next three years of government.
If that sounds bad, just navigate across to the ALP website, where the search function is broken (I don’t want to start drawing parallels to Satya Nardella’s leadership at this point because it just feels cruel).
Voters may not be demanding more detail on AI, but it is an unforgivable oversight to think that the latest advances in artificial intelligence will not start causing significant change to the Australian economy within the next term of government. Neither political nor business leaders can afford to miss this boat.
Of course, it’s easy to dismiss this analysis as one might do the perma-bear economist who has predicted six of the last four recessions. We’ve been hearing about the havoc AI was imminently about to wreak for years, why worry now? Well, to borrow a phrase, this time is different.
For the last couple of years, it looked like any company that built an application underpinned by a foundation model (e.g. Open AI) was not really an “AI business” it was just a wrapper. The thinking went that the best of these “wrapper” businesses would ultimately be subsumed by the foundation models, much like Google rendered other businesses obsolete in the previous phase of the internet. But the latest evolution of the foundation models has headed in a different direction – what’s called “System 2” thinking, which is where they conduct deeper, deliberate reasoning and move away from pre-trained instinctual responses.
The result of this shift is that foundation models can give better, more reasoned responses to queries, but the companies behind them are unlikely to move into specific sectoral domains – specific sectors are neither the strength nor the expertise of the model companies.
Instead, the opportunity is emerging for domain experts to build applications that leverage the models’ generalised reasoning power and apply them to specific fields – in other words, Generative AI-driven startups that provide services in any sector that is largely reliant on software for its provision (read: anything that’s not physical). This is an inversion of cloud computing’s SaaS: we are now talking about Service-as-a-software.
Take Harvey.ai as an example. The US company, founded in 2022, provides domain-specific AI for law firms, professional service providers and top US corporations. Law firms can now automate contract analysis, due diligence, litigation support, and regulatory compliance. A collaboration with OpenAI allowed the company to develop custom-trained case law models which enable complex legal reasoning. This means AI agents are now embedded into the daily workflows of Harvey’s clients – and tasks once undertaken by junior staff are replaced by the agent.
Versions of Harvey are emerging across a wide diversity of domains and jobs. Alongside lawyers, the work of software engineers, clinical scribes, personal assistants, customer support agents, and penetration testers are all already being affected by AI agents. These agents represent an enormous shift in the productivity and effectiveness of their users, but they undoubtedly also mean a reduction in the head count within the relevant firms.
We need to be clear-eyed about the implications of the latest AI advances. They offer extraordinary opportunity to a new breed of enterprising founders looking to address inefficiencies across most major industries. There is a generational chance to build global companies from Australia and our brightest minds should snatch it with both hands.
But we will also need insightful, agile policy to lead us harmoniously through the changes that will be wrought by this new technology. Big companies will be created, and new opportunities will exist for Australian founders to produce world-beating technology. But people will also lose their jobs, others will have to retrain, our infrastructure for supporting students and graduates will likely have to adapt.
Faced with such sweeping, imminent change, we cannot afford to step into the next election cycle in a policy vacuum. Nor can we afford to have our national debate dominated by boogey-man conversations about what AI will do. Both major parties owe voters a more considered approach to the AI revolution – like our best founders, we need to enthusiastically seize this opportunity whilst also managing its downside risks.
Bede Moore is the Global Chief Commercial Officer for Antler, where he oversees data and engineering. He is also a General Partner in the firm’s Australia Fund. He lives in Sydney.
Look back on the week that was with hand-picked articles from Australia and around the world. Sign up to the Forbes Australia newsletter here or become a member here.
