Australian workers want human oversight for AI autonomy

TeamViewer has published research showing that most Australian employees who use AI every day still want human oversight before it acts on their behalf. The findings point to a gap between routine use of AI tools and trust in more autonomous systems.

The survey found that 80% of respondents in Australia use AI at least once a day, while 64% do not want it to take independent action without a person involved. At the same time, IT leaders expect nearly 45% of digital workplace services in Australia to run autonomously by 2030, suggesting business ambitions may be moving faster than employee confidence.

Workers appear broadly comfortable with AI as an assistant, but become more cautious when software moves from generating outputs to making decisions or carrying out tasks on its own. Only 6% of respondents said they had no concerns about AI operating without a human in the loop.

That caution appears to carry a practical cost in the workplace. The research found that 65% of respondents often or always verify AI outputs before relying on them, spending an average of 2.34 hours a week checking AI-generated work. Another 55% said they do not always know when to trust AI and when to verify it.

Even so, the data does not point to a wholesale rejection of automation. Some 39% of respondents said they want AI to detect and fix issues early while keeping them informed, and 72% said they would be comfortable with AI acting autonomously as long as they could step in when needed.

Trust controls

Respondents identified several conditions that would make autonomous action more acceptable. Strong security and privacy protections were cited by 42%, while 40% wanted notifications before major changes and the same proportion wanted limits on what AI can access. Visible activity logs were identified by 31%, and 29% said rollback options would help build trust.

The findings suggest that visibility and control may matter as much as technical performance when companies introduce systems that can act without explicit prompts. The survey points not to a simple divide between using AI and rejecting it, but to a distinction between visible and invisible forms of autonomy.

IT stands out as the business function where attitudes appear most advanced. The study found that 32% of IT leaders described their departments as at least partially autonomous, and every IT decision-maker surveyed said they would be comfortable delegating at least one task to AI.

The tasks they were most willing to hand over were system updates, cited by 63%, device performance adjustments at 52%, and common device fixes at 45%. Those jobs are often repetitive and easier to measure, making them more likely candidates for early automation than areas where judgment is less structured.

Workers also appear to expect AI to change jobs more by reshaping tasks than by eliminating roles. By 2030, 36% of respondents said they expect AI to assist while they continue doing most tasks, and 26% said they expect to delegate more work and focus on higher-value activity. Only 2% said they expect their role to be eliminated.

The research was conducted among 4,200 respondents across nine markets, including Australia, the United States, the United Kingdom, France, Germany, Italy, Japan, Singapore and Brazil. Participants included managers and employees across a range of industries and company sizes.

Matthew Percy, Vice President of Global Digital Employee Experience at TeamViewer, said the central issue for employers is trust rather than outright resistance.

"The data shows that employees are not rejecting AI or the idea of greater autonomy - they are asking organisations to earn their trust," Percy said.

He said acceptance depends on people understanding what the systems are doing and retaining a way to intervene.

"People are comfortable with AI acting independently when they understand what it is doing, why it is doing it and can step in when needed. The path to an autonomous workplace is therefore not about removing people from the loop entirely, but building the transparency, control and guardrails that give employees confidence in when AI should act and when human judgment is still needed," Percy said.

Mei Dent, Chief Product and Technology Officer at TeamViewer, said the shift to autonomous systems requires organisations to define when software should act on its own and when it should wait for a person.

"For years, software has largely been designed around what people tell it to do," Dent said.

She said the next phase depends on clearer boundaries around automated actions.

"Autonomous systems change that model. The next step is designing systems that know when to act, when to wait, and when to bring people into the decision. That requires clear guardrails, pre-approved actions, risk-based oversight, and transparency into what the system did and why. Autonomy will scale when organisations can prove it works in specific, trusted use cases, then expand it with confidence," Dent said.