Making a claim at the Fair Work Commission? These are the new AI rules you need to follow

A record 44,039 Australians lodged claims at the Fair Work Commission between July 2025 and April 2026.

Those claims cover everything from unfair dismissal and underpayment disputes to bullying, sexual harassment and, most recently, arguments over the right to disconnect.

In late August, the commission declared it’s facing an “unprecedented growth in its workload” – linking a 70% surge in claims over the past three years to the growing popularity of artificial intelligence (AI).

A survey of more than 400 unfair dismissal applicants from earlier this year revealed about 40% had used AI, mainly ChatGPT. It also found “sycophantic or hallucinatory AI outputs may reinforce the applicant’s position, elevating their expectations” of winning their case.

That’s why, from October 20 onwards, the Fair Work Commission will enforce new rules on how to use AI and properly disclose it.

Here’s what you need to know about those rules, and why trusting AI too blindly could end up costing you money and your case.

What are the new rules?

Under new guidelines and updated forms, anyone who uses generative AI to help prepare a document lodged in a commission case will be required to:

• disclose if, and how, generative AI was used
• check all facts and legal evidence in your document are correct – and this checking needs to be done by a human, not another AI tool, and
• ensure any witness statement is based on the witness’s own knowledge and words, and is declared to be true.

Failing to comply with those new rules could see documents given less weight or being disregarded; you could be ordered to pay others’ legal costs; or the case could be dismissed.

The commission also warns people not to enter personal information about others – including confidential case information – into public AI tools such as ChatGPT, Claude, CoPilot or Gemini. Lawyers and even OpenAI chief executive Sam Altman have also warned about serious legal and privacy risks of doing so.

Professional representatives – such as lawyers, paid agents or human resources advisers employed by a party in a Fair Work Commission case – will be held to a higher standard.

In their documents, they’ll be required to include hyperlinks to all cited cases. Legal practitioners who fall short may be referred to their professional regulator.

Beyond the Fair Work Commission, more courts across Australia have also been giving clearer guidance on how to use AI, especially for people who want to represent themselves in court.

Read more:
An ‘AI legal team’ has won its first case. It’s a rare win for access to justice

When AI and people get it wrong

Just days before the Fair Work Commission finalised its new guidelines, there was a significant decision showing why they are so urgently needed, for everyone’s sake.

In August, Deputy Commissioner Michael Easton declared he was taking “the very rare step” of ordering a former Aldi employee to pay some of the supermarket’s legal costs, “because his unreasonable conduct caused his former employer to incur those costs unnecessarily”.

Why? Because his AI-driven claim for unfair dismissal was doomed from the start.

The store assistant lost his job three days short of the six-month minimum employment period that would have made him eligible to make an unfair dismissal claim.

Yet he still filed an application “obviously prepared using AI”, with chatbot notes left in (“Here is your final clean copy-paste version”).

The worker was warned by the commission in advance that his case was legally hopeless. But his AI kept arguing, so he persisted – and it ended up costing him $1,230, part of Aldi’s legal bill.

The commission didn’t blame the tool: Easton said AI “can assist litigants when used properly”. But in this case, it had been treated it as a “quasi-legal advisor”, while real-world human advice was ignored.

Read more:
More people are using AI in court, not a lawyer. It could cost you money – and your case

Broader lessons for AI use

Research has shown undisclosed “shadow use” of AI is running ahead of laws or many workplaces’ policies.

This leaves workers and organisations exposed to exactly the kind of unverified, but confidently wrong, output that ended up costing the former Aldi worker not just wasted time on a fruitless case, but more than $1,200 in costs.

The Fair Work Commission’s measured response — clear rules and user education, not a ban — is textbook good governance.

It’s a model any workplace or individual using AI can learn from too:

• be transparent – say when and how AI was used in your work
• final fact checks should be done by a person – not another chatbot
• and keep sensitive personal or confidential information out of public AI tools.

Guzyal Hill is a practising lawyer, but wrote this article in her role as a researcher working on AI governance and national uniform legislation.