Introduction: a new decision-maker

A manager is facing a difficult situation. An employee has raised a complaint – around the same time their performance has begun to slip. Unsure how to proceed, the manager turns to ChatGPT, enters everything they know about the employee, and asks what to do. The AI returns a list of options. One of them is dismissal, and the manager acts on it.

Months later, the matter is before the Federal Court of Australia in a general protections claim. The manager has filed an affidavit explaining the reasons for dismissal and is now in the witness box. The cross-examination is methodical rather than dramatic: the prior emails, the HR notes, the chat log, each tendered in sequence to show that the manager’s version of events cannot withstand scrutiny.

This is not a hypothetical. That is how these cases unfold.

It is the next frontier of employment law risk in Australia, and most businesses are not prepared for it. AI-assisted decision-making is colliding with Australia’s general protections framework, and that collision is not on the horizon. It is already happening.

Making sense of what is at stake means grappling with a complex legal landscape, and a doctrine Australian courts have spent years refining: the operative mind of the decision-maker.

A complex legal landscape

Anyone who has spent real time in this space knows that Australian workplace law does not sit still. It is a layered and interlocking framework built from federal statute, modern awards, enterprise agreements, state and territory anti-discrimination legislation, and a growing body of regulatory guidance on AI being written in real time.

For HR professionals and people managers, the practical reality is that decisions about employees (disciplining them, restructuring their roles, dismissing them) operate under legal obligations that are simultaneously significant and invisible until something goes wrong. And when it goes wrong, it does not go wrong quietly nor away from public scrutiny.

These are not just HR issues. They’re career-defining events

A general protections claim can expose a business to compensation orders, significant civil penalties, and the kind of reputational damage that does not disappear from a Google search. A discrimination complaint can trigger even more heightened public scrutiny.

A general protections claim turns on the intersection of these two elements. The employee must hold a workplace right, and the employer must take adverse action, but it is the point at which the two meet that creates the risk.

General protections: the three-step test

Employee holds a workplace right
Employer takes adverse action
If both are shown, the burden shifts – the employer, not the employee, must prove the prohibited reason played no substantial part

The Fair Work Act 2009 prohibits employers from taking adverse action against an employee because of a workplace right, and the protection is broader than most people assume. It applies whether the employee has exercised that right, declined to, or is proposing to, and it can be triggered even where a third party has exercised or proposed to exercise a right on the employee’s behalf. In practical terms, an employer can be exposed the moment a complaint is made, and in some cases before it is even formalised.

Adverse action is equally broad. For existing employees, it includes dismissal, injury in employment, altering a position to the employee’s detriment, and treating an employee differently from others. For prospective employees, it extends to refusing to employ someone or discriminating against them in the terms on which employment is offered.

Contrary to popular belief, an employee does not need to have been dismissed to bring a general protections claim.

What links the two is the employer’s reason for acting. The protection applies whenever a prohibited reason was one of the reasons behind the decision, and under section 360 it is enough that just one reason was prohibited, even where legitimate reasons sat alongside it.

Here is where employers consistently underestimate their exposure. In most claims the employee would have to prove why the employer acted, but the law reverses that. Once the employee shows they had a workplace right and that adverse action followed, the employer must prove the unlawful reason played no substantial or operative part in the decision. We call this the employer’s burden of proof.

Before the Court, the employer must affirmatively disprove the causal connection

To discharge that burden, the employer must generally call direct evidence from the decision-maker (the person or persons deciding to take the adverse action), positively establishing that the reasons for the adverse action did not include an unlawful reason. That unlawful reason need not be the sole or dominant reason. It must simply be a “substantial and operative” one. If the employer cannot discharge the onus, the Court must accept that the action, like a dismissal, was taken for a prohibited purpose.

The question a Court must answer is deceptively simple: why did the decision-maker do what they did?

This puts the evidentiary quality of the employer’s decision-making at the centre of the case. In most general protections claims, it is not a secondary concern. It is the whole contest.