The government wants to amend the Workplace Relations Act, in part to prevent claims based on artificial intelligence from interfering with the Fair Work Commission (FWC). However, documents released under FOI raise questions about whether evidence of such issues actually exists.
Related Articles Block Placeholder
Article ID: 333740
The new bill, called “Building Collaborative Workplaces,” was introduced earlier this month and is expected to pass the House of Representatives on Monday after only limited debate, without going through a Senate committee. of Australian Financial Review Reported.
The bill has attracted attention primarily for provisions that critics say would give unions more power. However, when Workplace Relations Minister Amanda Rishworth announced the reforms in a media release, she said the aim of the bill was to protect FWCs from the “increase in applications of artificial intelligence”.
With record numbers of claims coming into FWC’s desks, it is clear that the FWC’s workload is becoming unmanageable. However, the evidence claiming that AI is the cause seems thin.
“To date, while AI has been used, we have not seen widespread use of AI by stakeholders, with just over two-thirds of members surveyed estimating that AI has been used in less than 25% of the issues they have handled to date,” reads a set of talking points prepared for FWC general managers ahead of last year’s conference. The talking points, along with hundreds of pages of internal FWC communications, were published in April and are available on the Right to Know website.
“Nearly all members (91%) who responded believed they had encountered party use of AI. No survey participant said they had not encountered party use of AI…If we conducted the survey again in six months, that number would be closer to 100%,” the document said.
The issue also revealed that the FWC does not have an approved AI detection tool. This means that “estimations of stakeholder AI use from this study are based on what members observe, not what stakeholders disclose or see through AI detection technology.”
The FOI documents also include an email from July 2025 in which FWC officials asked colleagues to send them examples of “instances in which parties have used, requested the use of, or believe may have used artificial intelligence.”
An FOI officer working for the FWC admitted in an email published by Right to Know that he “could not find an answer” to that question.
That didn’t stop FWC from pursuing AI stories. In February, FWC chairman Adam Hatcher said in a presentation to the Victorian Bar Association that the commission’s work had been “significantly disrupted by the availability and use of AI tools”.
Hatcher told the audience that FWC’s workload is likely to increase by 70% by the end of this fiscal year compared to three years ago.
“The only reasonable inference that can be drawn, in my opinion, is that this is driven primarily by the increased use of AI tools by potential litigants,” Hatcher said. “The first clue that the increase in filings was AI-driven was the widespread use of AI-generated language in pending applications. Once you understand what this looks like, it’s very easy to identify.”
FWC officials estimated earlier this month that preliminary results from a new internal survey showed that “approximately 40% to 50% of people say AI was used in their applications to the commission.”
In response, Liberal senator Jane Hume asked, “So out of the 44,000 or so applications so far, you’re probably expecting around 20,000–” Mr. Yi Bright interrupted her and said, “I don’t know if you can guess that…This is a small survey that focused on this year’s applicants and respondents. There were about 240 or 250 responses to the survey. I would say it’s a suggestive study.”
Related Articles Block Placeholder
Article ID: 331931
If this situation sounds familiar, you’re probably thinking of Labour’s failed FOI reform. This reform was also supported by arguments that persistent and frivolous applicants were using AI to overwhelm the bureaucracy. In this case, the government had an example. For example, the eSafety Commissioner was inundated with around 600 FOI requests from Australians asking what data the agency handles on social media.
However, the government also called for FOI reform, pointing out that “offshore actors” were abusing the system for “illegal purposes”, but could not provide any evidence for its claims.
Crikey It called on the FWC and Rishworth’s office to provide additional evidence of AI use that may have gone undetected during the FOI process.
The FWC pointed to a statement made by the general manager in May, adding: “We have launched an external investigation into the use of AI in general protection and unfair dismissal matters. It is in these types of cases that we have seen the most significant increase in the number of applications.”
“While this research is in its final stages, early results indicate that a significant proportion of applicants are using AI in their applications to the Commission.”
A federal government spokesperson said: “The Chair of the Fair Work Commission has made it clear that AI is a key driver of the recent increase in workloads.
“Our bill includes a number of important measures that will help the commission better manage its workload, including strengthening its ability to dismiss frivolous complaints, vexatious complaints, or complaints with no prospect of success.”
This article was first published by Clicky.
