Dock workers demand 28-hour week as AI threatens jobs

On the waterfronts of Sydney, Melbourne, Brisbane and Fremantle, a familiar argument has returned in a newly sharpened form: who gets to pocket the gains when technology makes work quicker, safer and cheaper. This time the demand is unusually stark. Dock workers represented by the Maritime Union of Australia want a 28-hour working week, with no reduction in pay, as DP World presses ahead with plans to deploy artificial intelligence and automation across its Australian container terminals.

In the union’s framing, the claim is neither a perk nor a provocation, but a dividend. If algorithms, sensors and remote operations are to replace the strain and exposure of physical labour, the union argues, then shorter hours should be the price of consent. The Maritime Union of Australia has said DP World’s use of AI has put jobs “in the crosshairs”, insisting that technology “should be used to improve workers’ lives, not destroy them”. Behind the rhetoric sits a straightforward bargaining proposition: if the company insists on changing the nature of work, the workers want the benefits paid out in time rather than redundancy cheques.

DP World is not a marginal employer in Australia’s trade ecosystem. Based in Dubai, it handles about 40 per cent of the country’s shipping containers and operates terminals in the four major port cities. Any disturbance in its labour relations radiates through supply chains, affecting retailers waiting on stock, exporters chasing delivery windows and consumers already accustomed to price shocks. The union’s leverage comes from that strategic position, and from the reality that port operations remain, for all the talk of “lights-out” terminals, a tightly coupled system in which small delays can compound quickly.

The current working week for dock workers at DP World is reported to sit between 32 and 35 hours. A move to 28 hours would represent a material reduction in labour time, albeit one the union contends can be offset by the increased throughput promised by automation. It is a debate that rarely survives first contact with spreadsheets. Management will argue that a shorter week at unchanged wages lifts unit labour costs, and that the savings from automation are already being earmarked for capital expenditure, energy bills, software contracts and shareholder returns. The union’s counter is that those spreadsheets are precisely the point: if automation improves productivity, the arithmetic should allow workers to capture part of it without being forced into longer shifts, harder rosters or more precarious employment.

This dispute has been building for months in the technical corridors where modern industrial relations now live: safety documentation, consultation protocols and staged implementation plans. The union has, by the company’s account, held up DP World’s automation programme for about 18 months while proposals pass through safety checks and a mandatory nine-month consultation period. Such processes can look, to an impatient executive, like obstruction dressed as compliance. From the union’s perspective they are a necessary brake on a transformation that, once installed, cannot be bargained back.

What DP World hopes to introduce is not a single robot wheeled onto the quay, but a suite of technologies that reassign tasks away from people and towards systems. Sensor-driven vehicles would replace human-operated trucks, moving containers around a terminal with the kind of precision that machines manage well and humans manage only with training, fatigue management and luck. Remote-controlled rubber-tyre gantries and quay cranes would push operators out of cabs and into control rooms, changing both the skills required and the risks borne. The business case is familiar: fewer accidents, more consistent output, more predictable scheduling, and terminals that can run with less vulnerability to absenteeism, weather and industrial action.

The social cost, as the union sees it, is equally familiar. A report from the Maritime Union of Australia has suggested DP World’s automation plans could threaten up to 1,000 jobs. That figure is politically potent, even allowing for the ambiguity that tends to cling to automation forecasts. Companies often stress that roles will “evolve” rather than vanish; unions note that evolution has a habit of ending in attrition, contracting and fewer secure positions. In ports, where work has historically been a pathway to stable, well-paid employment for workers without university degrees, the fear is not just unemployment, but the replacement of a coherent occupational community with a smaller pool of technicians and a larger fringe of casual labour.

Negotiations are taking place on a terminal-by-terminal basis, with rollout scheduled to begin in Brisbane in February 2027. That incremental approach is a recognition that a port is not a single factory but a network of local practices, equipment configurations and labour arrangements. It also creates a laboratory effect: what is agreed in one terminal can become precedent in another. For a union, that is both an opportunity and a risk. Win strong protections in one site and you set a standard; lose them and the loss can travel just as quickly.

DP World was contacted for comment. The absence of a public rebuttal does not mean the company lacks arguments. It is likely to insist that Australian terminals must keep pace with overseas competitors, many of which have been automating for years. It may present automation as a safety initiative, pointing to the hazards of heavy machinery, fatigue and human error. It may also argue that if it cannot achieve the cost base and reliability demanded by shipping lines, investment will be directed elsewhere. The union will reply that competitiveness is a phrase too often used to make workers finance a transition from which they are excluded.

It is tempting to treat the confrontation as a simple rerun of earlier struggles between capital and labour, with AI merely the latest item in the long catalogue of innovations that displace jobs. Yet the AI element is not decorative. Automation in ports has typically been visible and mechanical: cranes, gates, guided vehicles. The newer generation adds algorithmic management, predictive scheduling and optimisation software that can invisibly reallocate work, measure performance and reduce the slack that makes a shift humane. Workers’ suspicion is shaped not only by the threat of fewer jobs, but by the prospect of work being reorganised by systems that cannot be negotiated with in the way a supervisor can.

In that sense, the union’s demand for shorter hours reads as an attempt to establish a principle before the technology becomes normalised. If a company can, through AI, run a terminal with fewer people, then those remaining can be asked to do more, cover more shifts, maintain more complex equipment, and accept greater responsibility for systems they do not control. A 28-hour week is a way of insisting that the direction of travel must include a tangible improvement in workers’ lives, not merely a promise that new roles will be “higher skilled”.

The history between DP World and the Maritime Union of Australia gives the argument an edge. The negotiations come a little over two years after the two sides reached an agreement that brought months of rolling strike action to an end. That earlier dispute was over pay and working conditions rather than automation, but it underscored how quickly a port conflict becomes a national story once ships begin to queue. It also demonstrated the union’s capacity to sustain pressure, and the company’s exposure when labour relations break down.

There is, too, a broader international context that both sides will be aware of, whether they cite it openly or not. The current dispute sits among a series of worker push-backs against AI-enabled automation. In 2023 Hollywood screenwriters undertook a five-month strike in part to protest against the use of AI-generated content in scriptwriting, seeking to stop studios from using machine output to undercut writers’ wages and status. The lesson carried beyond entertainment: even highly skilled workers, with cultural cachet, felt compelled to draw contractual lines before the technology settled into the background.

In 2024, dockworkers in the United States demanded a ban on automation of gates, cranes and lorries. The most sweeping demands were rejected, but workers secured protections against the introduction of fully automated equipment. That compromise is instructive. It suggests that employers can succeed in deploying some automation while conceding limits that preserve particular categories of jobs or require negotiated safeguards. It also hints at a future in which labour deals become increasingly technical documents, specifying what counts as “fully automated”, what human oversight must remain, and which tasks can be substituted by software.

Australia’s dock workers appear to be pushing the logic further. Rather than attempting to forbid automation outright, they are asking to be paid for it in time. That approach may prove more politically palatable than a demand to stop technology. Governments and the public can be unsympathetic to what looks like Luddism, especially when delays contribute to higher prices. A shorter week, framed as sharing productivity gains, can be presented as a social bargain: allow automation, but require it to deliver a measurable improvement in working life.

The economic stakes are not trivial. Ports are the joints of global trade, and automation is often sold as the cure for congestion, labour shortages and volatility in shipping schedules. But the distributional question is unavoidable. If AI reduces staffing requirements and improves throughput, the benefits can flow to shareholders, to shipping lines through lower fees, to consumers through marginally cheaper goods, or to workers through higher pay or fewer hours. The union’s demand is essentially a bid to fix the distribution before the gains are quietly allocated elsewhere.

There is also the question of what sort of workforce Australia wants in its critical infrastructure. A highly automated terminal can be efficient, but it may also be less resilient if it depends on a smaller number of specialists and a chain of software suppliers. If cyber security incidents, system failures or vendor disputes interrupt operations, recovery can be harder when fewer people retain the practical experience of running the terminal manually. Unions sometimes overplay this argument, but the last decade has offered enough examples of digital fragility to make it more than scaremongering.

For DP World, the challenge is to persuade workers, regulators and customers that automation is a managed transition rather than a sudden cull. For the union, the challenge is to make a case that does not read as a pure wage claim dressed up as ethics. The 28-hour week will, inevitably, be attacked as unrealistic. Yet it taps into a deeper unease: that AI will arrive as a management tool first, a productivity tool second, and a social benefit only if forced.

Much will depend on the architecture of the eventual agreement. A shorter week could be paired with commitments on redeployment, retraining, redundancy terms and the definition of which tasks must remain human-led. It could include guarantees on minimum staffing levels, consultation rights as new systems are introduced, and limits on the use of performance monitoring. It may also have to confront a practical question that unions sometimes avoid: if hours fall, who covers the work during peak periods, and how is that labour recruited without creating a two-tier workforce?

Brisbane, slated for the start of the rollout in February 2027, is likely to become the test case. If the parties strike a deal there that preserves dignity and security while permitting new systems, it will be cited as evidence that automation can be negotiated rather than imposed. If they do not, the fight risks becoming a template for a wider industrial backlash, not only in ports but across other sectors watching closely as AI moves from demonstration to deployment.