Australia sets up Office of AI amid trust concerns

The Australian Government has announced a new Office of AI within the Department of the Prime Minister and Cabinet, part of a broader push to shape the use of artificial intelligence in the national interest.

The decision has drawn a mixed but largely supportive response from technology vendors and workplace experts, who see both opportunity and risk for businesses and employees. Compliance, operational resilience and trust in leadership are emerging as central themes in how the new body could affect organisations.

Phil Beeston, ANZ Sales Leader at LogicMonitor, said the announcement provides long-awaited direction for organisations experimenting with AI while navigating uncertainty around governance and risk.

"The 'AI in Australia's interests' announcement is an important step towards helping Australian organisations adopt AI with greater confidence. Clear national direction on trust, governance and accountability will give businesses a stronger foundation to invest and innovate. The next challenge is operational readiness.

"As AI becomes more embedded in critical services and organisations deploy more autonomous AI agents, they need visibility into how those systems operate, the decisions they make, and the infrastructure they depend on.

"Governance doesn't end once AI is deployed. Organisations need to monitor AI to understand how decisions are made, identify issues early, and maintain trust over time.

"Trusted operational data and observability give both people and AI the context to make better decisions. Australia's AI ambitions are strong. However, long-term success will depend on pairing good policy with resilient digital operations that let organisations deploy AI safely, operate it with confidence, and adapt as the technology evolves," Beeston said.

Workplace and recruitment specialists are focusing on the implications for staff trust and communication. They argue national standards will only go so far unless employers change how they talk about AI with their workforce.

Sarah Donegan, HR and Recruitment Expert at Samuels Donegan, warned that many employees already question leadership decisions and see AI as another pressure point.

"Trust in leadership is already under pressure, and AI has the potential to either strengthen it or erode it further. The technology isn't the problem. It's how organisations introduce it.

"We're entering a period where employees are asking questions they've never had to ask before. Is AI monitoring me? Will it replace me? Is it making decisions about my performance or career? If employers don't answer those questions openly, trust will continue to decline.

"You can't regulate your way to trust. Trust comes from transparency, communication and bringing employees on the journey. Legislation can set the guardrails, but culture is what determines whether AI is embraced or feared.

"The organisations that succeed won't be the ones with the most sophisticated AI. They'll be the ones that explain why they're using it, how it will be used and what safeguards are in place. When employees understand the purpose, they're far more likely to embrace the technology.

"We're already seeing employees use AI every day without telling their employer. That tells us something important. Many people trust the technology more than they trust how their organisation will respond. Closing that gap should be the priority.

"Leaders also need to remember that uncertainty creates rumours. If organisations stay silent about AI, employees will fill in the blanks themselves, and those assumptions are rarely positive," Donegan said.

Donegan said pressures such as inflation, restructures and disputes over flexible work leave staff more sensitive to any change in technology or process.

"I think we're at risk of entering a period of even lower workplace trust. Employees are already dealing with economic uncertainty, constant organisational change and debates around flexible work. AI adds another layer of uncertainty. Organisations that treat AI as a technology project will struggle. Those that treat it as a leadership and change management challenge will build trust and get far better outcomes.

"The Government is right to focus on creating a social licence for AI and putting sensible standards in place. But trust won't come from regulation alone. It will come from employers being open about how AI is being used, what it means for people's roles, and how it will create better jobs rather than simply changing them," she said.

From the enterprise technology side, SUSE's regional leadership is emphasising regulatory flexibility and digital sovereignty as key tests for the Office of AI and any accompanying rules.

Ben Henshall, General Manager, Australia and New Zealand, at SUSE, said the new structure could support productivity if it balances safety, local control and innovation.

"Australia's newly announced 'Office of AI' could be a win for local businesses. But as with any new government body and its associated legislation, facilitating innovation safely, with sovereignty as a key principle, will be crucial to the country's long-term productivity.

"Compliance is already the biggest bottleneck for 43 per cent of enterprise AI projects, according to SUSE's recent research. The good news is that compliance hurdles don't have to stall innovation, provided businesses are built on open, flexible platforms and models that can adapt to new rules as they land.

"That flexibility matters even more given the pressures already facing the sector, including rising data centre energy consumption and growing hardware shortages. Linux-based open-source platforms remain the fastest way for ANZ businesses to optimise the hardware they already have, and they consistently prove the most efficient, with the lowest memory footprint.

"As government legislation inevitably evolves, sovereign models underpinned by open source will be the ultimate counter-lever for enterprises, keeping them compliant while preserving their freedom to choose where and how they run their data," Henshall said.