AI is flooding Australia's Fair Work Commission with fake claims
TL;DR
The Fair Work Commission is bracing for a 70% increase in claims above its three-year average, not because of mass redundancies, but because the friction cost of filing a legal claim has dropped to zero. Employees are generating ten-page unfair dismissal filings in fourteen seconds using ChatGPT. The claims are frequently full of hallucinated case law and legally impossible remedies, but your business still has to respond. Your employment contracts were drafted for the laptop era; they do not cover AI surveillance, algorithmic management, or data leakage through public chatbots, and that gap is now a live legal liability.
Why is the Fair Work Commission suddenly overwhelmed?
This is not a story about a sudden collapse in corporate culture. The Commission is not seeing more genuine disputes, it is seeing what happens when the barrier to filing drops to nothing. An employee who, in 2020, needed a lawyer and a consultation fee can now produce a comprehensive, aggressively worded legal claim from their couch at 11:00 PM for free.
The problem is the quality of what gets filed. Employment lawyers are politely calling many of these submissions "rubbish", claims that cite legal precedents that do not exist, seek remedies the Commission has no power to award, and demand ten years of back-pay in a forum that has a strict six-year statute of limitations.
The AI does not know the difference, and the employee using it certainly does not know the difference.
They just know the machine produced a very official-sounding document that makes them feel powerful.
What does it actually cost you when a fake claim lands on your desk?
Even a completely fabricated filing costs you real money. Your HR team still has to read every page. Your external lawyers still have to bill you to identify the hallucinated case law, untangle the legally impossible demands, and draft a formal reply explaining why the entire premise of the claim is void. You are paying thousands of dollars to argue with a hallucination.
And here is the element most business owners have not considered: there is no lawyer-client privilege for AI chat histories. If an employee uses ChatGPT to draft their claim and the other side requests discovery, those chat logs are fair game. Every prompt, every hallucinated response, every fabricated citation is sitting on a server somewhere, entirely discoverable. The employee thought they were getting private legal advice. In reality, they were building a paper trail, one that could be used against them, and potentially against you if your own AI usage has been equally careless.
How is the Fair Work Commission responding?
The Commission has recognised how dangerous this is becoming. It is pushing for new rules requiring applicants to sign formal declarations verifying that the facts in their claim are accurate and that the cases they cite actually exist. It is warning of cost consequences for unrepresented litigants who blindly rely on AI hallucinations. The Federal Court of Australia has already issued a Practice Note declaring AI-hallucinated evidence completely unacceptable.
Meanwhile, the legal profession itself is accelerating. Employment law firm Zed Law recently revealed they have automated approximately 80% of their own firm operations using AI. Legal research that previously took "a day to two" is now completed in "5 to 15 minutes." The lawyers acting against your business are using the same tools, with the verification skills to do it properly. The aggrieved employee at midnight does not have those skills.
Are your employment contracts already obsolete?
Almost certainly. Contracts drafted before generative AI became ubiquitous were built for a world where "technology" meant a company laptop, an email account, and a swipe card. They do not contemplate:
- AI analytics tools layered over existing software to measure employee productivity
- AI agents summarising employee communications for management reporting
- Algorithms recommending who gets promoted, who gets rostered for overtime, or who gets flagged for a performance improvement plan
- Employees pasting confidential client data into public chatbots to write a report
- Questions of who owns AI-generated output produced using company data
A generic confidentiality clause from 2018 does not protect you from a data breach via a generative AI platform. You need specific, targeted language governing which tools employees are permitted to use, what data they are permitted to input, and who owns the output.
See where AI fits in your business. Free.
A 45-minute audit. We map the highest-value automations and what they're worth in time and money. No pitch, no pressure.
What does workplace surveillance law have to do with AI?
More than most employers realise. In New South Wales, the Workplace Surveillance Act requires employers to give specific notice before monitoring employees' computer use. Historically that meant disclosing email monitoring or web browsing tracking. Today, if you layer an AI analytics tool over your existing software to measure productivity, or use an AI agent to summarise employee communications, you have almost certainly triggered those surveillance obligations. If your contracts and policies have not been explicitly updated to cover AI monitoring, you are walking into a compliance breach, one that looks nothing like the problem your original contracts were designed to prevent.
Why does algorithmic management create discrimination liability?
If an AI system is recommending who gets promoted, who gets rostered, or who gets flagged for underperformance, you are making employment decisions with an algorithm. If that algorithm produces a biased outcome, and many do, because they are trained on historical data that reflects historical biases, you are exposed to discrimination claims you may not even know you are generating. Your managers may not realise the "recommendation" they received was algorithmically generated rather than based on objective human assessment.
Explainability is no longer a technical nicety. It is a legal requirement that is rapidly crystallising across every jurisdiction. If an employee challenges a decision that was influenced by an algorithm and you cannot explain how it reached its conclusion, you are in serious trouble.
What are global regulators signalling?
The regulatory posture is consistent across every major market:
- UK: The Treasury Select Committee has published responses from the Bank of England and the Financial Conduct Authority on AI in financial services. Regulators are actively moving to designate major AI providers as "Critical Third Parties" subject to direct regulatory supervision.
- UK government: An open letter co-signed by two Cabinet ministers warned every business leader that frontier AI model capabilities are now doubling every four months. The Cyber Security and Resilience Bill is progressing through Parliament with an expectation that businesses of all sizes will need to demonstrate active AI governance.
- Singapore: The Infocomm Media Development Authority has proposed the world's first international standard for testing generative AI systems, building frameworks for benchmarking and red-teaming AI models to verify safety and reliability.
The message from regulators across the region is entirely consistent: the grace period for ungoverned AI use is over.
What to do this week
Audit your employment contracts immediately. Find every clause that touches technology, confidentiality, performance management, and data ownership. Ask whether it contemplates a world where employees use AI tools daily. If the answer is no, it needs redrafting.
Write an AI Acceptable Use Policy, now. This is the single most important document your business can produce in 2026. It is not optional. Cover which tools are approved, what data employees can input into any AI system, the difference between a secure enterprise-grade AI environment and a public chatbot that ingests everything they type, and the consequences of breach.
Train your team on shadow AI. Your staff are already using unapproved tools to make their jobs easier. Acknowledge it, explain the risk, and give them compliant alternatives. Leaving the policy on an intranet page is not training.
Document every AI-driven HR decision. If you use AI to screen CVs, monitor productivity, or flag underperformance, document exactly how those tools work, what data they use, and how decisions are made. Undocumented algorithmic decisions are a discrimination claim waiting to happen.
Prepare your HR team for AI-generated disputes. When a grievance or claim lands on your desk, your first question should be whether the document was generated by a human or a machine. Build a fast process to identify hallucinated case law before you spend thousands treating a chatbot's fantasy as a legitimate legal threat.
The Fair Work Commission does not care whether you have five employees or five thousand. The flood of AI-generated claims is not a future prediction, it is happening right now, and the spillover is heading straight for businesses that have not updated their legal foundations to withstand it.
Where to from here
Book a free 60-minute AI audit, we'll explore exactly what workflows are worth augmenting with AI.
Live with passion & AI,
Brett
Host a podcast? Have Brett on as a guest.
Straight talk on implementing AI in real SMEs, no jargon, plenty of receipts from the businesses we run.
Frequently asked questions
Can employees use ChatGPT to file an unfair dismissal claim in Australia?
+
Yes. There is no rule preventing it. The Fair Work Commission has seen a surge in AI-generated filings that cite non-existent case law and demand remedies the Commission has no power to grant. The filing cost to the employee is zero; the response cost to the employer is not.
Are AI chat logs discoverable in Fair Work proceedings?
+
There is no lawyer-client privilege for AI chat histories. If an employee uses ChatGPT to draft a legal claim, those chat logs, including every prompt and every hallucinated response, are entirely discoverable by the other side.
What is the Fair Work Commission doing about AI-hallucinated claims?
+
The Commission is pushing for new rules requiring applicants to sign formal declarations that the facts and cited cases in their claim are accurate. The Federal Court of Australia has already issued a Practice Note declaring AI-hallucinated evidence completely unacceptable.
Does the Workplace Surveillance Act cover AI monitoring tools in NSW?
+
The Workplace Surveillance Act requires employers to give specific notice before monitoring employees' computer use. Layering an AI analytics tool over existing software to measure productivity almost certainly triggers those obligations. If your contracts and policies have not been updated to cover AI monitoring, you are likely in breach.
How fast are lawyers adopting AI compared to employees?
+
Employment law firm Zed Law has automated approximately 80% of its firm operations using AI. Legal research that previously took a day to two is now completed in five to fifteen minutes. Lawyers, unlike aggrieved employees drafting claims at midnight, know how to verify the output.
What is algorithmic management and why does it create discrimination risk?
+
Algorithmic management means using an AI system to recommend who gets promoted, rostered, or flagged for a performance improvement plan. If that algorithm produces a biased outcome, which many do, because they are trained on historically biased data, you are exposed to discrimination claims your managers may not even know are being generated.
Are small businesses exempt from AI employment law risks?
+
No. The Fair Work Commission does not distinguish between five employees and five thousand. If an AI-generated claim lands on your desk, you are in the system. If your contracts are outdated and your policies do not cover AI, you are exposed regardless of size.
Brett is a four-time founder (Darra Tyres, Gladfish, EzyTrac, Anaboo) and the operator behind AIOS, Anaboo's AI Operating System. He writes from inside the build, installing AI in his own businesses first and reporting back what actually moves the numbers. Based between Singapore, the UK and Australia.
