It’s a pleasure to join you at this symposium on Seizing the Opportunities of AI While Protecting the Fair Go. I want to begin by acknowledging the Australian Council of Trade Unions, the Centre for Future Work, and the Centre for Employment and Labour Relations Law for bringing us together today.
The opportunity of AI
Some voices tell us AI means the end of work. We hear that robots will take every job, that plumbers and carers will be replaced by circuits and code. But the truth is more hopeful. Across advanced economies, the challenge is not too few jobs, it is too few workers. With ageing populations, shrinking birth rates and labour shortages, the demand for human expertise will remain strong.
The Productivity Commission has found that AI could underpin a wave of productivity growth in Australia. That is not just a number on a page, it is a reminder that with the right choices AI can help us lift living standards and sustain the fair go.
MIT economist David Autor has argued that AI is not just about rules and routines, it is about weaving information with tacit knowledge to support decision-making. If used well, AI can empower more workers, not just elite experts, to take on higher-stakes tasks. Nurses could handle more complex diagnoses. Technicians could supervise advanced manufacturing processes. Case workers could resolve problems that once demanded a lawyer.
This is what Autor calls restoring the 'middle-skill, middle-class heart' of the labour market. In plain terms: AI can help rebuild good jobs that pay fair wages and demand genuine expertise.
Lessons from history
We have seen this before. During the Industrial Revolution, artisanal expertise gave way to mass production. At first, conditions were grim: long hours, dangerous factories, children pressed into labour. But with the rise of unions and social reforms, new forms of 'mass expertise' such as machinists, typists and operators built a thriving middle class.
Later, in the computer age, expertise was reshuffled again. Computers empowered professionals but displaced middle-skill workers. The benefits went disproportionately to the top, while inequality grew.
The lesson is clear: technology alone does not guarantee fair outcomes. It takes strong institutions, above all unions, to ensure that workers are not left behind. Just as the labour movement fought for safety laws in factories and fairness in offices, unions today must help shape the AI age.
Unions and the AI future
Unions are already leading the way. Across industries, they are bargaining for agreements that ensure AI serves workers, not just employers. They are demanding transparency when algorithms are used in hiring or rostering. They are fighting for training and upskilling so that workers can use AI to do more interesting and higher-value work.
We saw this leadership on display at the Economic Reform Roundtable here in Parliament House last month. From 19 to 21 August, the ACTU played a terrifically thoughtful role in the discussions. They made the case that workers must be partners in shaping how AI is deployed, not passive recipients of decisions made in corporate boardrooms. That voice was vital, and it helped ensure that the national debate on productivity and fairness put working people at the centre.
The government responded. The Treasurer announced 2 immediate steps: developing an Australian Public Service AI plan, and accelerating work on a national AI capability plan. Those measures are about making sure AI is used for the common good, not just private gain.
Examples of AI improving work
Let me give some practical examples of how AI can improve the quality of work when introduced responsibly.
- Healthcare: AI tools can help nurses and midwives read diagnostic scans, track patient data and manage workloads more effectively. This does not replace clinical judgment, it strengthens it, giving health workers more time to care for patients.
- Aged care: AI can monitor residents’ movements to prevent falls, detect changes in eating or sleeping patterns, and alert staff before small health issues escalate. That supports carers to provide safer and more personalised care, while reducing stress and burnout.
- Construction: AI-powered safety systems can monitor sites in real time, spotting hazards before they cause accidents. Far from displacing trades, this helps ensure that workers come home safe each night.
- Logistics and transport: AI can optimise routes, cut wasted fuel and improve scheduling. That reduces fatigue for drivers, lowers costs for firms, and reduces carbon emissions for all of us.
- Public services: AI can simplify paperwork and speed up routine approvals, freeing up frontline workers to spend less time on forms and more time with the people they serve.
Each of these examples shows the same pattern: AI as a complement, not a substitute. A lever for expertise, not a replacement for it.
Policy choices ahead
Jobs and Skills Australia recently estimated that 87 per cent of current jobs are more exposed to augmentation by AI than to automation. That means most jobs will change, not vanish. The challenge is to manage that change so it makes work more rewarding, not more stressful.
But none of this is automatic. There are dystopian paths too. We know some countries are already using AI for surveillance, repression and union-busting. We know that if left unchecked, AI could be used to deskill jobs, monitor workers or entrench casualisation.
That is why we need 3 things:
- Voice – Workers must have a say in how AI is introduced. Consultation and collective bargaining must cover algorithmic management and technological change.
- Skills – Training must be a right, not a privilege. Every worker should have access to lifelong learning to use new tools, not be replaced by them.
- Fairness – AI must be used to raise standards, not cut corners. Whether in healthcare, aged care, logistics or construction, AI should improve safety, quality and service, not drive a race to the bottom.
These principles are at the heart of union campaigns here and abroad. They ensure that AI becomes a lever for rebuilding the middle class, not eroding it further.
A fair go in the age of AI
The Australian ideal of the 'fair go' means that prosperity is shared. It means that technology should serve people, not the other way around.
The risk of AI is not mass unemployment, but the devaluation of expertise. If every task is treated as generic, then no job is valued, and inequality soars. But if AI is harnessed to extend expertise, to allow more workers to exercise judgment, solve problems and take responsibility, then work becomes more rewarding and society fairer.
That is the choice before us. AI could be used to build a 'WALL-E meets Mad Max' world, where humans are sidelined and wealth concentrates in the hands of a few patent-holders. Or it could be used to lower the cost of healthcare and aged care, create dignified jobs, and give millions more people the pride that comes from meaningful work.
The difference will not be decided in Silicon Valley boardrooms alone. It will be decided here, in forums like this, where unions, researchers and policymakers come together to set rules and standards.
Conclusion
Let me finish with a simple thought. Work is not just a pay cheque. It is about purpose, dignity and community. For too many workers, that dignity has been stripped away by decades of rising inequality and casualisation. But with AI, we have a once-in-a-generation chance to restore it.
If we get this right, if unions continue to lead, if governments listen, if workers’ voices are at the centre, then AI can help us build an economy where expertise is widespread, middle-class jobs are plentiful, and the fair go is stronger than ever.
The story of technology is not fate. The future of AI is not written. It is ours to shape. And with the leadership of the union movement, I know we can seize its opportunities while protecting the fair go for all.
Thank you.