What the University of Sydney Strike Teaches Us About AI, Change and Psychosocial Risk

Skodel Newsletter: When a Pay Dispute Is a Psychosocial Risk Failure
Skodel Newsletter  ·  September 2026
When a strike is actually a psychosocial risk failure
In this newsletter
1.  Is this just an industrial dispute? (plus the consultation duty in full)
2.  University of Sydney — the NTEU strike over AI
3.  Other common trigger events to watch for
4.  Hazards at play
5.  How to use this in your workplace
6.  Example controls
Key Figures
90 / 330
New positions the university promised in its last agreement that still haven't been filled
0
Enforceable AI protections currently written into the enterprise agreement itself
1. A common question
Isn't a strike over AI just an industrial relations issue?
Not only that. Job security and organisational change are named psychosocial hazards in every state's code of practice, and consultation on a change to systems of work is a standalone WHS duty, not just good bargaining practice. Whether a dispute plays out through an enterprise agreement or a HR project, the same test applies: did you consult before deciding, and did you leave people with anything certain in writing?
Quick reference  ·  the consultation duty in full
Sections 47–49, WHS Act 2011 (NSW) require a PCBU to consult, so far as reasonably practicable, with workers directly affected by a WHS matter — expressly including proposed changes to the workplace, plant, systems of work, or ways of doing work. Consultation means sharing relevant information, giving workers a reasonable opportunity to express views before the decision is made, and genuinely taking those views into account.
The model Code of Practice on psychosocial hazards lists job security and poor organisational change management as recognised hazards in their own right, to be identified, assessed and controlled with the same rigour as any physical hazard — regardless of whether the change is a restructure, a merger, or a new technology rollout.
2. Skodel Case Analysis University of Sydney — NTEU strike, 2 September 2026
Hundreds of University of Sydney staff held a 24-hour strike on 2 September, picketing every Camperdown campus entrance. Alongside long-running claims on job security and workload, the NTEU's headline demand was new: enforceable, contractual protections against AI being used to cut jobs, intensify workloads, or override professional judgement. "This fight is about much more than protecting our own jobs," said NTEU NSW secretary Vince Caughley. "It's about protecting what a university is."
The University's counter is that AI governance belongs in policies that can be updated as the technology evolves, not locked into the enterprise agreement — a position that, from a psychosocial risk lens, keeps the employer's flexibility intact while leaving staff with nothing certain in writing. That's the same gap that shows up in the 90 unfilled positions above: promises made outside an enforceable instrument are easy to make and hard to hold anyone to.
Read the full breakdown →
Let's look at some potential controls for a hazard like this.
Job insecurity
Put commitments in writing, in the instrument that actually binds you  ·  give a specific timeline for filling promised roles, not an open-ended one.
Poor organisational justice
Share the draft, not the finished framework  ·  build in a genuine feedback window before the decision is locked in, and show what changed as a result.
3. Other Common Trigger Events
Most workplaces won't face a strike, but almost every workplace will face one of these moments: predictable, foreseeable, and worth the same proactive consultation step.
Rolling out AI or new automation
Change management
Enterprise agreement renewal or bargaining
Job insecurity
Mergers, outsourcing or restructures
Organisational justice
New staff monitoring or productivity-tracking software
Intrusive surveillance
Same principle as the case above: if the event is predictable, the consultation step should be too, not left to whichever project team happens to be running it at the time.
4. Hazards At Play
Job Insecurity
Promised roles and protections left outside any binding instrument
High
A major technology and role change introduced without agreed safeguards
High
Poor Organisational Justice
Framework shared after it was finalised, not shaped by genuine consultation before
Mod-High
5. How To Use This In Your Workplace
Treat "we'll keep this in policy so it can flex" as a flag to check whether staff actually wanted certainty instead.
Build a genuine consultation checkpoint into any AI, technology or restructuring project before the plan is finalised, not as a step after sign-off.
List the predictable change events on your own calendar — EA renewal, tech rollouts, restructures — and check each has a documented consultation step attached, not just a comms plan.
6. Example Controls
For the trigger events above. Starting points, not a checklist — evidence from your own people should always come first.
Poor organisational change management
Consult staff before changes are finalised, not just inform them afterward  ·  give adequate notice and transition time, with support during the change period.
All 17 hazard categories, with practical controls for each.
Upcoming session  ·  18 September 2026
Psychosocial Hazards in the Workplace: From Compliance Obligation to Practical Action
With Sam McKenzie, Ian Fagan and Ben Chinwah, hosted with Benchmark OHS.
Register for the webinar →
Skodel · Psychosocial risk assessment for Australian workplaces · skodel.com