By EMA Manager Employment Relations & Safety Paul Jarvie
When cinematographer Halyna Hutchins was fatally shot and director Joel Souza injured on the set of the film Rust, attention quickly turned to the people involved.
Who handled the firearm? Who checked it? Who pulled the trigger? Who should have known there was a live round in a revolver being used as a prop?
These are understandable questions after a tragedy. But from a workplace health and safety perspective, there is another question that matters just as much: how did the systems surrounding those people allow a single mistake, or a series of mistakes, to end in death?
The shooting happened during a rehearsal in New Mexico in October 2021. The state’s Occupational Health and Safety Bureau subsequently investigated the workplace conditions that contributed to the fatality and the injury to a second person.
It would be comforting to view Rust as a uniquely American event, created by circumstances that could not arise here. Unfortunately, New Zealand has already had its own warning.
It has happened here
In April 2016, two students received serious throat lacerations during a production of Sweeney Todd at an Auckland high school. Real straight razors had been adapted for use in a scene simulating the cutting of two characters’ throats.
WorkSafe found that the school had failed to implement adequate systems for identifying the hazards associated with the production. It also identified inadequate controls, including the absence of an effective system for inspecting and tracking the razors, and insufficient training and instructions for the students using them.
The injuries could have been avoided by using theatre props instead of real blades. The case is a stark reminder that a demand for realism can introduce unnecessary risk. If an audience cannot tell the difference between a real blade and a safe prop, why introduce the sharper object in the first place?
The same principle applies well beyond stage or screen. Whenever a business chooses a process, tool or material with the potential to cause serious harm, it should ask whether there is a safer way to achieve the same result.
Look beyond the person
When an incident occurs, people are naturally drawn towards blame. We look at the last person who touched the equipment or made a decision and treat their action as the cause. That’s hindsight, not prevention.
People make mistakes, particularly during familiar and repetitive work. Routine helps us operate efficiently because we do not need to consciously reconsider every step. But routine can also make a changed condition harder to notice.
A worker may complete the same check correctly hundreds of times. On the next occasion, something is different. A guard has shifted, a component has been replaced, an item is in the wrong location or information has been misunderstood. The worker sees what they expect to see and continues.
If the safety of the entire process rests on that person noticing the difference, the system has little capacity to tolerate normal human error. The better question is not simply, “Who failed?” It is, “What failed, and why did the system not catch it?”
Build layers, not single points of failure
Where the potential consequence is severe, one check is rarely enough. For example, a firearm can be unloaded by the person responsible for it, physically marked to show it has been cleared and independently inspected before anyone moves into a potentially exposed area.
Each step creates another opportunity to identify a problem. If one person misses something, another control should prevent that error from reaching the point where somebody is harmed.
This principle can be applied to machinery isolation, confined-space entry, hazardous substances, vehicle movements, lifting operations and many other critical risks.
Businesses need to identify where an ordinary mistake could produce an extraordinary consequence. They can then introduce what are sometimes called “error traps”: controls designed to detect an error before it travels through the system.
Two people glancing at the same document and agreeing may not provide a meaningful double-check. The second check must genuinely test the first rather than simply confirm an assumption.
Test whether the controls work
Having a process on paper is not the same as knowing it works. Controls can become routine too. Workers may complete a checklist without actively looking for a changed condition. Supervisors may assume a task is safe because it has been performed many times without incident.
Over time, repeated success can normalise risk and create confidence that the system cannot fail.
That’s why businesses need to test their controls. In a safe and carefully managed setting, this could involve simulations, drills or introducing a harmless variation to see whether it is detected. The purpose is not to catch workers out. It is to understand whether the system works when the situation differs from what people expect.
Fresh eyes can also be valuable. Rotating reviewers, involving workers from another team or asking someone unfamiliar with the routine to observe the task may reveal assumptions that have become invisible to those doing the work every day.
Workers themselves need to be part of this process. Asking them to describe how a task is normally performed, rather than how the procedure says it should be performed, can expose workarounds, time pressures and practical gaps in existing controls.
Plan for safe failure
Health and safety often concentrates on stopping failure altogether. That remains the preferred outcome, but it should not be the only plan.
Modern vehicles provide a useful analogy. Safety systems aim to help the driver avoid a crash, but designers also recognise that prevention will not always succeed. Seatbelts, airbags and crumple zones are there to reduce the consequences when it does not.
Workplaces need the same mindset. If a worker presses the wrong control, what prevents the machine from moving unexpectedly? If a load is not secured correctly, what keeps people away from the fall zone? If hazardous material is released, what contains it? If one inspection is missed, what other barrier remains?
The aim is to allow a process to fail safely, or at least to ensure that the consequences are far less severe.
The lesson for New Zealand businesses
The lasting lesson from Rust is not confined to firearms or the film industry. It is about designing work on the understanding that people are fallible, circumstances change and controls can weaken over time.
Deliberate unsafe conduct should be addressed but most incidents are not caused by someone knowingly choosing a harmful outcome.
Businesses will learn more by examining the conditions surrounding an error than by stopping the investigation at the person closest to it.
That means using foresight rather than relying on hindsight. Identify where mistakes are likely, build independent layers of protection, test whether those protections work and ensure a single error cannot lead directly to catastrophic harm.
When something goes wrong, resist the immediate urge to ask who failed.
Start by asking what failed, why it failed and what needs to change so that the next mistake has a safer ending.