Artificial intelligence has begun its descent from its hype peak. The breathless talk of super-intelligence and job wipe-outs is yielding to a sober reality now. Even OpenAI chief executive Sam Altman concedes we may be in a bubble, and thoughtful AI critics like Karen Hao are gaining prominence.
This global reality check is a gift to Australia. Ours is a country allergic to techno-utopian grandstanding. Faced with Silicon Valley’s evangelism and Europe’s caution, the average Australian instinctively asks: what’s it actually do, who gets hurt, and who’s paying?
That is precisely the pragmatic mindset we need.
The challenge is not to legislate based on science fiction or panic about mass unemployment, it is to treat AI as a tool and use it in a way that improves our lives.
The federal government’s National AI Capability Plan, due by end of year, is where the rubber hits the road, and where we must see a pragmatic strategy come to life.
The Business Council of Australia’s Accelerating Australia’s AI Agenda provides a clear, business-backed blueprint to inform the plan, and the government should be adopting its simple case: AI can lift productivity, sharpen national security, and strengthen our sovereignty if we regulate it sensibly.
Done poorly, regulation is a brake. Done well, it is a catalyst. Our objective for this plan must be a trusted regulatory landscape that signals stability to investors, provides certainty to businesses, and gives Australians confidence in the systems touching their lives.
Singapore offers a clear model for how to align AI governance with national capability-building, combining strict accountability frameworks with heavy investment in skills, infrastructure, and local innovation. Australia could adopt a similar approach.
The core principles are clear: focus on use and impact, not technology. Regulate outcomes, not algorithms.
That means we must give regulators like ASIC and ACCC the clarity and resources to interpret existing laws through an AI lens. We must scale our rules based on real (not imagined) risk: light-touch for benign uses, proper scrutiny for anything touching health, money or justice.
The conversation also needs a reset.
AI is not the miracle disruptor sweeping away jobs wholesale, nor the sentient threat of Hollywood movies. It is becoming infrastructure that is deeply entangled with how organisations function.
AI is increasingly embedded into workflows rather than being just a standalone tool. It takes the first cut of analysis, drafts reports, or optimises logistics. The productivity benefit is not in replacing humans, but from freeing them up for higher-value work in judgement, creativity and client engagement.
That means leaders should stop asking “which jobs will be lost?” and start asking “how do we redesign roles so machines and people better complement one another?”
Australia also needs to abandon any delusions of going it alone in AI. This is a global game of capital, hardware, and talent. It’s time to formalise an AI partnership with the United States and bake it into our National AI Capability Plan.
Australia should not copy the US regulatory approach but should aim for interoperability. We have to be able to work well with the global driver of this innovation.
AUKUS is a template. If we can share nuclear submarine technology, we can share AI governance and safety research. This creates an innovation bloc that strengthens economic resilience and national security.
An AI system certified as safe in one jurisdiction should be recognised in the other. Supply chains for semiconductors and compute power should be secured collectively. Australia’s recent critical minerals agreement with the
The US demonstrates how partnerships can capitalise on each nation’s comparative advantages.
Today, a handful of US firms control most frontier compute. Australia should negotiate deals for cheaper national compute time now to ensure long-term access and competitiveness as demand and costs increase.
The global AI conversation is shifting from the sugar high of hype to harder questions about value, risk and integration. Australia should embrace that pragmatism.
The National AI Capability Plan must reflect this. Less prophecy, more policy. Less fixation on job loss, more on workforce redesign and better reskilling.
The hype may be fading, but the hard work is only beginning.
Mike Bareja is director, digital technologies, AI, cyber and future industries at the Business Council of Australia, and Helen Zhang is the co-founder of Intrigue Media
Do you know more? Contact James Riley via Email.