What’s holding Australia back from seizing AI boom?

Mr Rubinsztein asserts a “Darwinian element” to AI adoption, where companies must adapt or be left behind in a rapidly accelerating technological landscape.

“I truly believe it’s AI or die,” Mr Rubinsztein said.

And Nuix is in survival mode. In the past year its shares have more than halved, with the company now valued at $960.5m, as it reels from a tough four years involving high-profile governance and regulatory woes that have increased its cost base and weakened profitability.

Mr Rubinsztein has been steadily seeking to turn around the company’s fortunes, rebuilding it to become more AI-focused, including developing its own model rather than rely on OpenAI or Anthropic. This differentiates itself in a market dominated by hyperscalers, Microsoft, Amazon and Google.

Mr Rubinsztein said that if every human decision within an organisation is made even 1-3 per cent better through AI, the entire “organism” of the business will “respond and adapt better and faster to the environment.” This isn’t just another technology, he stressed, but a fundamental shift requiring leadership to “understand, to use, to embrace and understand the boundaries of possibility that AI gives you.”

He was speaking ahead of Nuix’s XLR8 (accelerate) conference, which will be held on Tuesday and feature former Optus chief executive Kelly Bayer Rosmarin and former NSW Minister for Customer Service and Digital Government Victor Dominello.

While he advocates a survival of the fittest mentality towards AI, the widespread adoption of the technology is not without its significant hurdles. Ms Bayer Rosmarin said the real challenge isn’t merely technological readiness, but rather “psychological and organisational readiness for transformation.”

She cautioned against isolated experiments, stressing the need for a “transformation mindset” that rethinks entire processes. For example, she said that simply giving coders an AI assistant won’t revolutionise software development if the underlying workflow remains unchanged, as developers spend less than half their time “hands on keyboard.”

“You actually have to rethink the entire life cycle of software deployment and transform that entire life cycle,” Ms Bayer Rosmarin said.

“If you’re doing tiny, isolated experiments, you’re not going to hit on the use cases that are powerful and transformative enough that you want to scale them.

“Now I wouldn’t be disparaging of Australian companies. They are experimenting, they are learning. They’re taking the time to get their data in order, which is absolutely necessary. But there is a missing piece in the ecosystem.”

Mr Dominello said the public service remained risk averse – even more so in the aftermath of the ‘Robodebt’ scandal.

He said this “human system failure” has made the public service even more cautious, leading to “fear paralysis” but it was critical for the government to embrace AI to spur greater adoption across the economy.

“About one in every five people employed by government, we’re about 30 per cent GDP. We we regulate pretty much everything from smuggler arms to sleep like in your life. So unless we tackle that element, it’s really, really hard for industry to sort of move forward,” Mr Dominello said.

But he said Australia’s “goat track” data architecture compared with countries like Estonia and Singapore, which boast “crossroads” for safe data sharing, remained a stumbling block.

He argued that government silos, rather than technology, are the barrier to efficient AI integration, leading to “Jurassic” paper customs declaration cards for international arrivals and a lack of the ”frictionless experience” that a digital identity could provide. “It’s madness,” he said, describing the continued use of outdated paper systems at airports.

Mr Rubinsztein acknowledged the inherent risks in AI and data management but insisted that adoption is “not optional.” He warned that “training your AI on bad data just simply means that you make stupid decisions faster.”

“The leadership challenge isn’t simply employing a human to own it, but it’s actually getting the leadership to understand, to use, to embrace and understand the boundaries of possibility that AI gives you.

“That itself becomes a kind of an iterative process, where when you understand what it can do, you then understand how you can use it. You can understand how you can embed it in processes.”

This article first appeared in The Australian as What’s holding Australia back from seizing AI boom as other countries leap ahead