Dr Maria Hameed Khan
Centre for Decent Work and Industry
Queensland University of Technology

Dr Maria Hameed Khan reflects on the important themes discussed at the AI symposium convened by the Centre of Future Work and the Business and Human Rights Centre, RMIT University.

The use of artificial intelligence (AI) is no longer a futuristic concept in Australian workplaces. AI and algorithms are actively managing rosters, monitoring performance, overseeing redundancies, and making recruitment and selection decisions. While vendors selling AI technology promise efficiency and productivity to organisations, the reality is often increased worker surveillance and poorly implemented technology that negates any benefits to managers or staff and, in fact, undermines decent work conditions.

On the 9 April I attended the symposium Which way forward? AI and Decent Work. The key insight I took from the symposium is that, as businesses race to adopt AI in Australian workplaces, our regulatory frameworks that are meant to protect workers are lagging. We are currently facing a critical disconnect: the consultation processes meant to protect workers are triggered far too late, occurring only after decisions to implement AI are already in motion. Simultaneously, the ubiquity of public-facing generative AI (GenAI) tools is reshaping worker behaviours through unofficial, or shadow AI use.

A race without rules

Technology has always shaped the nature of work, but the current pace of AI adoption is unprecedented. Globally, there has been a sharp rise in AI-powered tools to automate tasks, support complex workflows, and aid decision-making. All this while also increasing employee monitoring, a trend mirrored in Australia. Yet, our policy landscape remains inadequate. Recent government inquiries, such as the 2025 Productivity Commission inquiry into harnessing data and digital technologies, have largely focused on the impact of digital technologies on economic productivity while sidestepping the nuances of human costs and labour market implications from the adoption of these technologies.

Currently, AI adoption in the workplace is viewed as a procurement issue – a transaction between software vendors and management – rather than a catalyst for shifting workplace relations. This means Australian organisations buy opaque systems without independent evaluation of accuracy, safety, or decision-making impacts. Constrained by regulatory compliance rules, procurement decisions are made based on projected productivity and efficiency gains. Because there are no standardised metrics for impact, and ongoing risk management is minimal, work conditions are consequently eroding.

The failure of consultation

The risks of AI adoption are obscured by vendor-promised efficiency. There is also an emerging trend of ‘AI-washing‘, where companies use AI-enabled productivity to rebrand pre-planned job cuts. The most glaring evidence of this failure is the Commonwealth Bank (CBA) case. In a striking irony, workers were directed to train the very AI systems intended to replace them. The result was not efficiency, but a breakdown in service, increased customer aggression and the failure of AI agents to handle complex, nuanced queries. This case highlights how the current industrial consultation processes can become box-ticking exercises and underscores the urgent need for stronger protections. Moreover, this gives rise to worker resistance driven by fears of future job loss.

Future directions

The presentations and panel discussions at the symposium confirmed that the path forward on AI in Australian workplaces requires a fundamental shift from reactive compliance to proactive bargaining to protect worker interests and agency. Unions must start incorporating AI into collective agreements now, establishing clear boundaries around surveillance, data usage, and the right to refuse automated tasks.

Legislative reform must also follow. We need a digital duty of care, where the burden of proof for AI safety lies with the employer, not the employee. There is also a need to rethink when consultation occurs as AI technology continues to be introduced into the workplace.

Ultimately, work must be well-paid, dignified, and co-determined. If we allow AI adoption and algorithmic management to continue unchecked, we risk reducing workers to mere inputs in a data-driven system. Therefore, enforceable, structural change is needed to ensure decent work conditions in the age of AI.

This article is part of a suite of papers written following the symposium ‘Which Way Forward? AI and Decent Work’. Read more here.