Fair Work Claims Are Up 70% in Three Years. AI Is Why.

The Fair Work Commission processed 44,039 lodgments between July 2025 and April 2026 — already matching the full-year record of 44,075 set in 2024-25, with two months still to run. Commission President Justice Adam Hatcher told the Victorian Bar Association in February that the tribunal's total workload has grown more than 70 per cent in just three years. He pointed to a single primary driver: generative AI.

Since ChatGPT launched in November 2022, the barrier to filing an unfair dismissal or general protections claim has collapsed. A $89.70 filing fee, a few minutes with a chatbot, and a form lands on the Commission's desk. Unfair dismissal claims are up 41 per cent between 2022-23 and 2024-25. General protections claims are up 62 per cent. The Commission projects 50,000 to 55,000 lodgments for the current financial year.

Justice Hatcher demonstrated the problem himself. He fed a basic hypothetical dismissal scenario into ChatGPT and received a filing-ready form — complete with a fabricated witness statement — in under ten minutes. The chatbot predicted compensation of $15,000 to $40,000 for a claim with no reasonable prospects of success.

The actual median payout for general protections conciliated settlements sits between $4,000 and $5,999. Less than 1 per cent of claimants receive a formal judgment award. But the person filing doesn't know that, and the chatbot won't tell them.

The quality of AI-generated submissions is a separate problem. In Riley v Nuvei Australia [2026], an applicant cited case law that didn't exist — pure AI hallucination. General-purpose chatbots produce incorrect legal information 69 to 88 per cent of the time, according to legal AI benchmarking research cited by Fair Work Workplace Solutions. Even specialised legal AI tools hallucinate at rates of 17 to 43 per cent.

A response, a legal review, management time, emotional bandwidth. With 85 per cent of dismissed employees now contesting their termination — up from 76 per cent before the AI surge — the odds of facing a claim have increased materially for every Australian employer.

The pattern is worth understanding. Two-thirds of general protections applications now come from people ineligible for unfair dismissal, per Fair Work Workplace Solutions analysis of Commission data. This suggests applicants are using AI to reframe their claims and bypass eligibility requirements. The filings look more sophisticated in form, if not in substance.

For a trades business with five to twenty staff, or an accounting practice with a small team, defending even a meritless claim can mean thousands in legal fees and weeks of distraction. When the volume of claims doubles, so does the expected cost of employing people.

On 24 March 2026, the FWC published draft guidance requiring anyone who uses generative AI to prepare documents for lodgment to disclose that fact. Commission forms will include a dedicated 'Use of GenAI' section. Failing to disclose could result in the document being given reduced weight, disregarded entirely, a costs order, or case dismissal.

The disclosure rule won't reduce the volume. But it gives the Commission and employers a tool to test credibility. If someone can't explain the legal principles their filing relies on, that tells the member something. The Albanese Government has also signalled legislative reform to streamline case management, following Justice Hatcher's February call for action. But reforms take time. The volume is here now.

First, invest in your termination process. Section 387 of the Fair Work Act sets out what the Commission considers when assessing a dismissal: was there a valid reason, was the person notified, were they given an opportunity to respond, were they warned about unsatisfactory performance? A few hundred dollars spent on a proper procedure saves thousands defending a claim you should never have attracted.

Second, document everything. Written warnings, performance conversations, policy breaches — recorded contemporaneously. AI-generated claims often fabricate or embellish facts. Your best defence is a real-time record that tells the actual story.

Third, don't use AI to manage your own termination process without legal review. AI-drafted dismissal letters miss enterprise agreement requirements and create procedural vulnerabilities that claimants will exploit. The irony of the AI-claims surge is that employers using AI carelessly on their side of the table face exactly the same quality problems.

Key takeaways

Want to know where your employment processes are exposed?

AI has changed the cost of filing a claim. A short conversation can identify whether your termination procedures, documentation, and HR workflows are ready for the new volume. Book a call.