AI Is Changing the Psychosocial Risk Profile: Are Your Managers Equipped to Spot the Signs and Respond?

Artificial intelligence is changing how work is done. As organisations move quickly to adopt new tools, automate tasks and find productivity gains, there is another question that deserves attention:

What happens to the people whose work is changing?

AI is not a distinct category of psychosocial hazard. However, the way it is introduced can introduce recognised hazards such as high job demands, low job control, lack of role clarity, poor organisational change management and poor organisational justice. Safe Work Australia: Psychosocial hazards

For employers, it’s important that AI adoption is not treated purely as a technology project. It is also critical that work design, change management and people risk are at the front of implementation considerations.

The risk may not be AI itself

For Bronwyn Dennis, Manager, Learning Solutions at iHR Australia, the starting point is to look at AI in much the same way as any other significant workplace change.

“Any changes that we make in the workplace, whether that's a technology change like AI, a new system of work or a new tool, can have a psychosocial risk component.”

The technology may be new, but the underlying principle is not. Employers need to understand what is changing, identify potential hazards, put appropriate controls in place and then check with the people doing the work whether those controls are actually working.

One of the challenges in getting buy-in is that the people leading an AI implementation and the people experiencing it are often at very different stages of the journey. Senior leaders may have already spent months discussing the technology, considering its intended benefits and thinking about what it means for the organisation. Employees may hear about the tool for the first time at the point when the decision has already been made.

Their questions are based on being at the start of the change and could include:

• What does this mean for my role, and will it change?
• Will my responsibilities change?
• Will I need new skills, and if I do, what training will I receive?
• What happens if the new tool gets something wrong?

Bronwyn says leaders can sometimes mistake this uncertainty for resistance to change, when employees may simply be at the beginning of processing something leaders have been considering for months.

AI may shift the demands rather than reduce them

One of the biggest assumptions surrounding AI is that if technology makes something faster, workload must go down. That is not always what happens.

AI may remove routine work while leaving employees with a greater concentration of complex tasks. Time saved creating something may be replaced by time checking outputs, validating information, correcting mistakes and applying professional judgement to determine whether the output is fit for purpose.

As Bronwyn puts it:

“What looks like a small change can completely reshape the work landing on that person’s desk.”

That means employers should be cautious about immediately translating AI efficiencies into higher performance expectations.

It’s not just about whether AI saves time. It’s also worth looking at what it changes about the job itself: the pace, the expectations and the demands on the person doing it.

A useful assessment looks beyond the time saved on an individual task and considers the pace, complexity and accountability of the job.

The same applies to job control. AI may give an employee a useful tool that makes their work easier, or it could narrow their discretion through automated decisioning, which is difficult to challenge. Technology can also influence how work is allocated, how performance is measured or how decisions are made.

Understanding that difference requires organisations to listen to the people doing the work.

Managers need to notice what changes in the work

Psychosocial risk management works best when it is preventative.

When absenteeism, engagement scores or other organisational measures indicate a problem, an individual may already have been struggling for some time, and managers may have missed an opportunity to address the problem early.

That makes the relationship between managers and their teams critical so a manager can identify early:

• An engaged employee becoming withdrawn.
• Someone who previously managed their workload comfortably beginning to work excessive hours or make unusual mistakes.
• An employee stopping participation, missing breaks or simply saying they feel overwhelmed.

None of those behaviours automatically means AI is the cause. However, they are reasons to ask.

Bronwyn describes the manager's role through a simple approach: Notice. Ask. Act.

Notice when something changes.

Ask early enough to understand what is happening.

Act where something about the work is contributing to the problem.

That action might involve clarifying expectations, reviewing workload, providing more training, increasing support or escalating something the manager cannot resolve themselves.

Managers should then check whether the response worked.

What if managers do not have the answers?

This may be one of the hardest parts of AI-driven workplace change.

Uncertainty about future roles can be especially difficult when an organisation speaks about efficiency without explaining what will happen to the work. An employee wants to know what AI will mean for their role, and the manager doesn’t know yet.

Bronwyn says managers should avoid two tempting responses: offering reassurance they cannot guarantee or avoiding the conversation altogether.

“We don’t want to promise something we can’t guarantee or pretend we know more than we do. But we also can’t just go quiet because we don’t have the answer yet.”

Sometimes the right response is simply:

“I don't have the answer to that yet. I understand why you're asking. I'll find out what I can, and we'll keep talking as we know more.”

Managers do not need to know everything. Credibility comes from being clear about what is known, what remains undecided and when employees can expect an update.

Consultation cannot be an afterthought

The people doing the work will often see consequences that an implementation team cannot.

They know whether AI has genuinely reduced workload or simply changed it. They know whether a new process has created duplication, removed useful autonomy or made part of their job more difficult.

That is why consultation matters before, during and after implementation.

As Bronwyn explains, employees may not describe what they are experiencing in the language of psychosocial hazards or risk controls. But they can tell you what has changed in their work.

For employers, it is also important to consider any applicable consultation obligations when introducing significant workplace change. Fair Work Ombudsman: Consultation and cooperation in the workplace

The question leaders should be asking

AI will continue to change Australian workplaces. Some tasks will become easier; roles will evolve and expectations will change.

The organisations that manage this well will not necessarily be those with the most sophisticated technology. They will be the ones that recognise that introducing AI means changing work, and changing work means considering the people doing it.

So the question for leaders shouldn’t simply be, “Where can we use AI?” It should be, “If we do use AI here, how will it change the way our people experience their work, and are our managers equipped to recognise and respond to that change?”

That is where AI implementation becomes a psychosocial risk issue.

iHR Australia works with organisations to strengthen how they identify, assess and manage psychosocial risks in practice. Our services include Managing Psychosocial Risks training for managers and leaders, psychosocial risk audits and reviews, and practical support to assess existing approaches, identify gaps and strengthen controls.

If your organisation is reviewing how workplace change, including AI, may be affecting psychosocial risk, we can help you build a practical approach that supports both compliance and effective day-to-day management.