AI is flooding Australia’s employment system, forcing a rethink of how law is practiced

Australia’s employment law system is bracing for a sharp rise in claims driven by widespread use of generative AI, as workers turn to chatbots for cheap, instant legal help – and HR teams and employers grapple with the fallout.

Carly Stebbing, head of employment law at LEAP Legal Software, said recent Fair Work Commission guidance reflects how seriously regulators now view AI-generated applications.

“The recent guidance out of the Commission has suggested they’re expecting a 70% increase in claims as a consequence of Gen AI,” she said. “Some of the rubbish that they’re getting filed… completely overwhelmed a HR team with baseless allegations just constantly generated, you know, by whatever [the applicant] was using in the background.”

In one recent matter, Stebbing noted, the Commission was so critical of an AI-padded application with “no prospects of success whatsoever” that it took the rare step of calling for an application for costs against the self‑represented applicant – a clear signal that simply pasting prompts into ChatGPT and filing whatever comes out carries real consequences.

Commission pushes ‘human verification’ and costs risk

Far from banning AI tools, the Fair Work Commission’s emerging approach is to insist that humans remain accountable for what’s filed in their name.

Stebbing said proposed guidance would require applicants to make a declaration that the facts in their claim are accurate and that any cases they cite actually exist and support the propositions they’re being used for.

“They understand [these tools] exist and people are going to use them,” she said. “But they’re saying, you’re still responsible for checking its output.”

For a jurisdiction that has not traditionally been “a cost jurisdiction”, Stebbing believes the new focus on declarations is significant. If someone blindly relies on hallucinated cases or remedies, “there may be cost consequences” – even for unrepresented litigants.

She worries, however, that the Commission may be asking more of everyday workers than is realistic: “Asking the average person to check that that case exists and stands for the legal proposition… is a step too far. An average person probably can’t do that.”

The practical effect, she hopes, is that more people will pause and seek proper advice – from community legal centres, the Commission’s Workplace Advice Service, or a lawyer – rather than treating a chatbot as a free, infallible barrister.

AI‑written claims: Empowerment or false hope?

On one view, AI has democratised access to legal language: employees who once felt powerless can now generate detailed claims at the click of a button.

But Stebbing said that in employment law, generic large language models are already leading vulnerable workers astray.

She recalled acting for an employer in a case where an unrepresented worker had clearly relied on ChatGPT. “They brought an unfair dismissal application and they were seeking general damages… You can’t do that in that jurisdiction. It is not available,” she said.

The same applicant claimed ten years of alleged underpayments in a forum that only allows six years’ recovery because of limitation periods, and asn’t even the correct jurisdiction for that type of wage claim.

“Do I think that helped that person? No, I don’t,” Stebbing said. “That person would have been better assisted by going to their community legal center… or using the Fair Work Commission’s workplace advice service.”

She also warned that workers confiding in chatbots have no confidentiality or legal professional privilege:

“People are using it as like a quasi‑psychologist… Something that they’re not remembering, too, is that none of that is confidential. You don’t have lawyer‑client privilege there, which means that all of that, your prompts and the responses, are discoverable by the other side.”

As costs applications become more common, she predicts litigants may find themselves ordered to hand over their AI chat histories – the digital equivalent of giving up their private diary of the dispute.

Legal tech versus ‘general’ AI: Why the platform matters

Despite the risks, both Stebbing and Zed Law director Nandan Subramaniam are adamant: when properly designed and embedded, AI can be a game‑changer for employment law.

Stebbing drew a sharp line between generic AI tools and specialist legal tech integrated into law firm systems.

A legal research task that once took a junior lawyer “a day to two” can now be generated to a similar standard within “five to 15 minutes of massaging and prompting,” she said.

“Our real skill as lawyers is not really the legal research,” Stebbing argued. “It’s in collating that and then exercising judgment, especially in a space like employment law, where there are so many… differences in fact patterns between cases.”

Tools can also summarise large volumes of matter documents – the “enormous stack” every litigator knows well – so lawyers can walk into client meetings with a fast, accurate brief and easily hand over files to colleagues.

Crucially, all of this happens within the “safe and privileged and confidential environment” of a firm’s own practice management platform, rather than in a consumer app on someone’s phone.

Contracts written for email, not AI

Subramaniam says the law around workplace technology is badly lagging the reality of AI‑driven workplaces.

“Most employment agreements in Australia were drafted for a world where ‘technology’ meant a company laptop and email,” he said. “Now employers are deploying AI across hiring, performance monitoring and workflow automation, but the contractual and regulatory framework hasn’t caught up.”

That disconnect is already showing up in disputes.

“We’re seeing employers expose themselves to risk because their employment agreements don’t contemplate AI‑driven surveillance, algorithmic decision-making, or the IP implications of employees using tools like ChatGPT on the job,” Subramaniam said.

The more employers lean on AI to make or inform decisions – from recruitment to discipline to redundancy – the more critical it becomes to update contracts, policies and notices.

‘Don’t wait for Parliament’: Fixing the AI gap now

Subramaniam said the savvier employers aren’t sitting back and waiting for bespoke “AI in the workplace” legislation.

“The employers getting this right aren’t waiting for legislation,” he said. “They’re updating their employment agreements to address AI tool usage, data ownership and monitoring, and they’re being transparent about it.”

Existing law already bites. In NSW, the Workplace Surveillance Act requires notice for computer surveillance – an obligation that can easily be triggered when employers layer AI analytics or monitoring tools on top of current systems.

“Employers layering AI on top of existing systems without updating their disclosure obligations are walking into a compliance issue,” Subramaniam warned.

On the employee side, he says, education is just as important: “If you paste client data or proprietary information into an AI tool without your employer’s knowledge, that’s a potential breach of your employment obligations. The legal education needs to go both ways.”

Law firms as AI ‘laboratories’

At Zed Law, AI isn’t just something the firm advises on – it’s how the business runs.

“We’ve automated roughly 80% of our firm’s operations using AI, from client intake and document drafting to CRM management and invoice processing,” Subramaniam said.

That hands‑on adoption shapes the firm’s advice to employer clients. “We’re not theorising about what AI adoption looks like,” he said. “We’ve lived the implementation, the privacy considerations and the change management. That practical experience matters when you’re advising an employer on how to roll out AI responsibly.”

For HR leaders, that lived experience can be critical as they try to harness AI for efficiency while staying on the right side of workplace law, privacy and discrimination rules.

HR under pressure: ‘Do more with less’, but safely

For HR professionals, AI presents a double‑edged sword. On one side, generative tools have fuelled a flood of claims, many of them poorly framed or hopeless. On the other, legal‑grade AI – properly deployed – can help stretched teams cope with rising volumes and complexity.

Stebbing said Commission forecasts of a 70% increase in applications such as unfair dismissal and general protections over the three‑year average should be a wake‑up call.

“If there is this overwhelming increase in employment law claims, it only makes more of a case, in my view, for utilising the legal tech tools that are available to practitioners in the most efficient and productive way,” she explained.

But that must happen inside secure, enterprise‑grade environments – not via ad‑hoc use of public chatbots that may compromise confidentiality and privilege.

“Everyone’s trying to do more with less,” Stebbing noted. The firms and HR teams that will cope best, she argues, will be those who:

• Adopt legal‑specific AI into their existing practice or HR systems

• Keep confidential data within protected environments

• Re‑focus their human effort on strategy, judgment and negotiation – the parts of employment law that machines cannot yet replicate

As Australia’s employment law landscape shifts from “digital first” to “AI first”, both Stebbing and Subramaniam agree on one thing: AI isn’t going away – but neither is the need for human lawyers and HR professionals at the centre of the process.