A Macquarie University computing academic has used artificial intelligence to win a historic case at the Fair Work Commission (FWC), with the technology beating out his employer’s human lawyers.

The commission ruled earlier this month that Greg Baker, a casual academic at the Sydney institution, was entitled to a permanent part-time position based on new casual conversion laws introduced two years ago that had not yet been tested.

It’s an even more historic result due to the use of AI by the applicant.

‘For the lulz’

Baker has been a casual employee since 2023 at Macquarie University’s School of Computing, where he has taught the same subject each semester and was scheduled teach it again this year.

When he workshopped some career planning with OpenAI’s ChatGPT Pro platform he spends $US200 ($280) on per month, the AI tool said he may be able to use new workplace laws to transfer to a part-time position.

“[ChatGPT] didn’t actually mention that nobody had successfully run a case on it,” Baker told Information Age.

“It told me I should do it and followed through with it – it gave me all these arguments and made sure it all happened.”

Baker’s application to move to a part-time position was allegedly refused by the university, so he lodged the proceedings with the FWC using AI tools.

“Part of me just did it for the lulz – it’s pretty funny running a case with ChatGPT and seeing what happens,” he said.

“Nobody was more surprised than me when ChatGPT won.”

To make his case, Baker said he fed case material, emails, conversations, and meetings records to the generative AI platform, and trained his AI agents to check their own work for accuracy.

“The very interesting part is how rapidly it got better,” Baker said.

“By late last year it was pulling apart the university’s arguments very, very skilfully and finding all sorts of interesting precedents in all sorts of ways.

“It was doing very in-depth work, reading through hundreds of Fair Work rulings and finding neat analogies.”

Baker’s victory over his employer was first reported by the Australian Financial Review.


Greg Baker has taught as a casual employee at Macquarie University’s School of Computing since 2023. Image: Shutterstock

Casual or part-time?

The case hinged on wording in the Fair Work Act which requires casual work to be without a “firm advance commitment to continuing and indefinite work”.

The FWC looked at whether the university could realistically decline to offer the work to Baker, if the work would continue being available to him, and whether staff who were permanent workers did the same job as him.

It found that Baker’s work was “more akin to recurring fixed-term periods of part-time employment, rather than genuine casual employment”.

He won the case against his employer and said it showed the power of AI in democratising the legal system.

But the FWC has recently warned that it is being inundated by AI slop which is “significantly disrupting” its operations, with a sharp uptick in its cases and inaccurate submissions.

“It’s a really fraught problem,” Baker said.

“On the one hand this case would never have happened if I’d been forbidden from using AI, and we’d still be in limbo over whether the legislation does anything.

“But at the same time, it’s very expensive for Fair Work, with self-represented litigants using ChatGPT but doing it really badly and wasting everyone’s time.

“I don’t know the solution.”

Democratising access to justice

While much of the reporting around the use of AI in law has focused on hallucinations and inaccurate AI filings, Baker suggested errors are becoming less common.

“Hallucinations are much, much rarer than they were one or two years ago, and you can just make sure you’ve got the rules in place,” he said.

“It’s inevitable that AI will get better, and within a few years the stupid stuff we’re seeing at the moment will disappear – but that won’t reduce the caseload.

“It’ll be ChatGPT very accurately pointing out problems in institutions’ behaviour.”

Easier and cheaper access to justice would help open up the legal process and make people more aware of their rights, he said.

“In computing we have spent the last decades putting people out of work in different job roles, and now finally we have this situation where we’re finally able to give something back to the non-computing community,” Baker said.

“Here’s AI, it can bring about access to justice, it can change the balance of power between institutions that exploit workers.

“Finally, after tech being seen as the bad guys always making bad jobs worse, now there’s the opportunity to show that what tech creates can be empowering and bring about much better outcomes.”