
Australia’s Workplace AI Guide Makes Worker Consultation the First Step

On 12 November 2026, Australian workplace relations minister Amanda Rishworth released a draft best-practice guide for adopting AI at work, with public feedback due by 27 January 2027. Rishworth put its central idea plainly in that release: “AI works best when the workforce is at the centre of the adoption process.” The draft asks employers to involve affected workers and their representatives before deployment changes how jobs are done.
The guide is voluntary, so it does not give workers a new right to approve or veto an AI rollout. It follows discussions with employers, small businesses, unions and academics. Guardian Australia’s launch-day reporting describes the dispute over optional guidance and the draft’s concerns about intrusive surveillance, unrealistic performance standards and heavier workloads. Those risks help explain why worker input appears before job redesign in the proposed sequence.
What happens before an AI rollout

The guide groups its principles under Plan, Enable and Embed. Selecting a use case and assessing workplace readiness come before cooperative planning and worker engagement. Consultation is therefore the first people-facing step once an employer has identified a proposed use, rather than a request for workers to assess an unspecified technology. The later principles cover job design, skills, workforce transitions, governance and safety.
For a small online business or a team working across locations, that sequence translates into a set of decisions to make before AI becomes part of ordinary work:
- Define the use case. Identify the task the system would support and the result the business wants. A proposal to use AI for a particular task gives affected staff something concrete to assess: the work involved, the decisions they make and the exceptions they handle.
- Assess readiness. Consider whether the proposed use fits the current process and whether workers have the time, skills and oversight needed to use it. Access to a tool alone does not show that the surrounding work is ready to change.
- Engage affected workers and representatives. Explain the proposed change while staff can still influence its design. Workers can identify where an extra checking step would add work, where an AI output would need human judgment and which parts of the job the proposal overlooks.
- Redesign jobs and handoffs. Specify what the system would do, what remains a human responsibility and who handles errors or unusual cases. Automating a routine task can leave people with more complex work, so the design needs to account for their full workload.
- Build skills and plan transitions. Match training to changed duties, including how workers assess outputs and respond when they are unsuitable. If tasks move substantially, the draft’s workforce-transformation principle points beyond a one-off introduction to the tool.
- Establish governance and manage risks. Decide who is accountable for the system’s use and how its effects on work will be reviewed. That includes working conditions and safety as well as whether the system performs its intended task.
This is a practical sequence drawn from the draft’s principles, not a prescribed procedure for every employer. Its central dependency is clear: workers need to know what is proposed before they can describe its likely effect, and their account needs to reach the people designing the change.
Why consultation can change the design

A faster task does not necessarily mean a lighter job. If AI produces more work for a person to check, or adds another step before they can finish an existing task, the total workload may rise. Rishworth warned at a Sydney business forum that introducing AI without proper worker consultation could have that effect. Asking workers about the whole process can reveal costs that a measure of the automated task would miss.
Monitoring creates a related design choice. An AI system may generate information about how quickly people work or how often they intervene. If managers use that information to set performance expectations, the consequences extend beyond the tool’s stated purpose. The draft’s concerns include surveillance, poor rostering, discrimination and intensified standards; consultation gives affected staff a chance to describe how those uses would play out in their roles.
The effects may also differ across a workforce. The guide raises concerns about younger and older workers, First Nations workers, women and people with disability. Changes to entry-level tasks, training opportunities or performance measures can matter differently to each group. Involving workers before deployment makes those differences part of the job-design decision while the proposed use can still be altered.
Where the voluntary guide meets existing duties
The draft recommends cooperation throughout AI adoption, but existing workplace rules determine whether consultation is legally required for a particular change. The Fair Work Ombudsman’s consultation guidance explains that legislation, awards and enterprise agreements can require employers to notify affected employees and their representatives, discuss significant workplace changes and consider the views raised.
Award consultation clauses generally apply when an employer has decided to introduce a major change to technology that is likely to affect employees significantly. Consultation may also be required for changes to regular rosters or ordinary hours. Under the major-change clauses described by the Ombudsman, considering workers’ views does not mean obtaining their consent. The applicable obligation depends on the workplace’s award or agreement and the effects of the proposed change.
That distinction matters on both sides of a rollout. The draft encourages an employer to seek worker input early enough to shape an AI use and the jobs around it; an applicable consultation clause sets the enforceable requirements for a qualifying change. Public submissions remain open until 27 January 2027, after which the government can revise the guide before deciding its final form.
Also read:
Related articles


Cal.com’s Free Plan Beats Calendly Until a Team Needs Routing

Ways to Improve Resource Allocation Across Business Departments

Clockify or Toggl Track? Lower cost trades off daily polish

Gallup finds U.S. worker engagement below 30% as AI strains clarity

How Freelancers Can Build a Repeatable Video Service with AI
Subscribe to our newsletter
Get the latest Web3, AI, and crypto news delivered straight to your inbox.