Australia’s Fair Work Commission has announced a review of its processes to cope with what it described as an estimated 70% workload increase over three years, driven in part by the proliferation of generative AI assistance tools. The commission, which handles unfair dismissal claims, wage disputes, discrimination, bullying, and workplace sexual harassment, said the surge is directly affecting its ability to provide timely dispute resolution, according to a statement published on Friday.
The numbers tell the story. The commission received 44,039 lodgments between July 2025 and April 2026, with two months still remaining in the financial year. The full 2024–25 year saw a record 44,075 lodgments. The commission is on pace to exceed that record by a significant margin.
How AI changes what gets filed
The commission attributed the increase to several factors: more people representing themselves in workplace cases, budget constraints, resourcing challenges, and the spread of generative AI tools that make it easy to produce polished-sounding but often generic content. The implication is that AI is lowering the barrier to filing a claim, enabling people who might previously have decided a case was not worth pursuing to generate a detailed submission in minutes.
Generative AI tools, such as large language models, can draft legal documents, letters, and arguments with minimal user input. This capability is particularly appealing to individuals who cannot afford legal representation—a group that includes many workers in low-wage industries. However, the ease of use comes with risks. AI-generated content may cite statutes or precedents that do not exist, misinterpret jurisdictional rules, or include irrelevant details that confuse the issues. The Fair Work Commission already reported instances of fabricated case law and vastly inflated damage claims, forcing adjudicators to sift through pages of material to separate fact from fiction.
The commission published draft guidance in March requiring anyone who uses generative AI in preparing documents for lodgment to disclose that fact. The guidance warned that AI-generated information may be incomplete, inaccurate, or fabricated. A new “Use of GenAI” section will be built into all commission forms. This move is intended to alert decision-makers to the potential unreliability of submissions and to encourage filers to verify AI output.
What the commission is doing about it
The response includes trialling a new system in which senior staff help parties try to resolve disputes informally earlier in the process, before cases consume full hearing time. This early intervention approach aims to reduce the number of applications that proceed to formal hearings, thereby easing the burden on commissioners and support staff. The commission has also reviewed how it manages applications and is considering deploying an AI voice agent to help triage calls to its helpline.
The irony of a tribunal overwhelmed by AI-generated filings considering an AI tool to manage the influx is not lost. Australia has already backed AI in other parts of its legal system, including a government-supported chatbot that helps splitting couples divide their assets. But the logic is sound: if generative AI is increasing the volume of inbound work, automated triage may be the only way to keep pace without proportional increases in staffing that budget constraints have already ruled out.
Additionally, the commission is exploring the use of machine learning classifiers to flag potentially frivolous or AI-generated claims early. This would allow staff to prioritise cases that require urgent attention while routing others to alternative dispute resolution channels. However, such tools must be carefully designed to avoid bias against legitimate self-represented litigants who do not use AI.
Not just an Australian problem
The pattern is emerging across the Tasman as well. Radio New Zealand reported last month that tenants in New Zealand are using AI to support applications to the Tenancy Tribunal, creating extra work and backlog. In one case, a tenant used AI to file a claim for $40,000 over issues including unsafe drinking water and a broken dryer. The tribunal awarded $80.
The case illustrates a recurring problem with AI-generated legal filings. The tools can produce arguments that sound authoritative but cite legal principles that do not apply in the relevant jurisdiction, claim damages wildly out of proportion to the harm, or reference legal frameworks from other countries entirely. Adjudicators then have to work through pages of material to identify which parts are relevant. In New Zealand, the Tenancy Tribunal has reported a 25% increase in applications over the past two years, with staff noting a clear correlation with the rise of freely available AI chatbots.
Similarly, in the United Kingdom, the Employment Tribunals Service has observed a spike in self-represented litigants using AI to draft claims. While official data is not yet available, anecdotal reports from tribunal judges suggest that many of these submissions are poorly structured and contain legal errors. The UK’s Ministry of Justice is reportedly studying the Australian approach to disclosure requirements.
The financial complaints parallel
The Australian Financial Complaints Authority, which handles disputes in financial services, told Bloomberg it has also seen increased AI use in how consumers engage with financial firms and lodge complaints. A spokesperson acknowledged that AI can help some people articulate their concerns, but warned that AI-generated complaints “can sometimes include irrelevant, inaccurate or generic information, or may use legal arguments that don’t apply in Australian law.”
AFCA said it encourages people to keep their complaints simple because lengthy AI-generated submissions slow down the resolution process by forcing staff to work through large volumes of material to identify the actual issues. The advice amounts to an admission that more words do not mean a better case, and that AI’s tendency to produce verbose, confident-sounding output is actively counterproductive in dispute resolution.
The AFCA received over 80,000 complaints in the 2024-25 financial year, a 15% increase from the previous year. While not all of this growth is attributable to AI, the authority estimates that approximately 10% of complaints now show evidence of generative AI use. This percentage is expected to rise as the technology becomes more embedded in everyday life.
Access to justice or access to noise
The tension at the heart of the issue is genuine. AI tools can democratise access to legal processes for people who cannot afford lawyers, a point governments and AI companies have promoted as a public benefit. But when the same tools generate filings that are longer, less accurate, and harder to process than what a human would produce unaided, the net effect may be to slow down the system for everyone.
Proponents of AI-assisted legal filings argue that the technology can help level the playing field. For instance, a worker facing unfair dismissal may not know the correct legal terminology or how to structure a claim. An AI chatbot can guide them through the process, ensuring all necessary information is included. However, critics counter that the quality of AI output depends heavily on the quality of the underlying model and the user’s ability to fact-check the results. Many users, lacking legal expertise, are ill-equipped to verify AI-generated content.
The Fair Work Commission’s disclosure requirement is a step towards transparency, but it does not solve the underlying problem of volume. If the commission receives 50,000 AI-assisted claims per year, even if each is flagged, the sheer number will still strain resources. The early intervention pilot and the AI voice agent are attempts to manage the flow, but sceptics wonder whether these measures will be enough.
There is also a broader concern about the environmental and ethical implications of generative AI. Training and running large language models consumes significant energy and water. If the primary use case for these models becomes generating low-quality legal filings that clog tribunals, the societal cost may outweigh the benefit. Some legal scholars have called for stricter regulation of AI in legal contexts, including mandatory accuracy audits and liability for harm caused by negligent AI use.
A changing landscape for dispute resolution
The experience of Australia’s Fair Work Commission is likely a harbinger of what other jurisdictions will face. As generative AI tools become more sophisticated and widely available, any institution that accepts written submissions from the public will have to adapt. The challenge is not just about handling volume, but also about maintaining the integrity of the decision-making process.
Some tribunals are experimenting with their own AI tools to assist in reviewing submissions. For example, the UK’s HM Courts & Tribunals Service is developing a natural language processing system that can summarise key points from lengthy filings. However, deploying such systems raises questions about bias, transparency, and the potential for errors. If an AI triage system misclassifies a legitimate claim as frivolous, the consequences for the claimant could be severe.
Another approach is to educate the public about the limitations of AI in legal contexts. The Fair Work Commission’s guidance is one example, but many observers argue that more proactive measures are needed. Legal aid organisations could offer workshops on how to effectively use AI as a drafting tool while recognising when to seek human advice. Similarly, AI platform developers could incorporate disclaimers and fact-checking prompts into their interfaces.
The financial services sector is also grappling with the issue. Some banks and insurance companies report receiving AI-generated complaints that are so generic they could have been written for any company. This makes it difficult for the business to understand the specific grievance and respond appropriately. In response, some firms have started using their own AI tools to analyse incoming complaints and route them to the correct department, but this again raises the spectre of a technological arms race.
Ultimately, the debate over AI in dispute resolution reflects a broader societal challenge: how to harness the benefits of generative AI while mitigating its downsides. The Fair Work Commission’s workload crisis is a concrete example of what happens when adoption outpaces adaptation. Other tribunals, both in Australia and abroad, will be watching closely to see what solutions emerge.
The commission plans to report back on its process review by the end of 2026. In the meantime, it has urged all users of its services to “keep it simple” and to verify any AI-generated content before submission. Whether this advice will be heeded remains to be seen, but one thing is clear: the era of AI-assisted legal filings has arrived, and institutions must evolve or be overwhelmed.