The robots were supposed to replace the lawyers, but first they're replacing the paralegal filter that kept bad cases out of the system.
The Summary
- Australia's Fair Work Commission is on track for a 70% workload increase over three years, driven partly by people using ChatGPT and other AI tools to file workplace disputes.
- The tribunal's solution: deploy AI to handle the AI-generated caseload, including automated processing and AI voice agents for helplines.
- This is the first documented case of a legal institution openly planning to use AI to filter AI-generated work, a pattern that will ripple across every sector where gatekeepers face asymmetric workload.
The Signal
Since ChatGPT launched in 2022, people have increasingly turned to AI for legal assistance, lowering the barrier to filing cases. Australia's Fair Work Commission now faces a flood. Murray Furlong, the Commission's general manager, says the rise of generative AI tools coincides with more people representing themselves, creating a perfect storm of volume without corresponding increases in quality or merit.
The math is brutal. A 70% workload increase over three years means the tribunal is drowning. Furlong notes this is happening alongside budget constraints and resourcing challenges, the exact moment when institutions can least afford a surge. The AI democratization dream meets the reality of finite institutional capacity.
"These impacts, taken together, are having a direct effect on the Commission's ability to provide timely, efficient and effective dispute resolution services to the community."
Here's the uncomfortable truth: AI didn't make people better at filing legitimate cases. It made filing cases easier. There's a difference. When you remove friction from a process, you don't just get more of the good stuff. You get more of everything. The tribunal is learning what every spam filter learned two decades ago: accessibility without quality control is a vector for noise.
The Commission's response is to fight fire with fire, deploying AI for automated processing and voice agents on helplines. This is the first institutional admission that the solution to AI-generated work is AI-powered filtering. Every customer service department, every HR team, every regulatory body is watching.
Key implications:
- Legal institutions face asymmetric workload: AI makes it easier to create cases than to process them
- The "AI for access to justice" narrative now includes the cost: institutional capacity strain
- First-mover institutions will establish patterns for AI-filtering-AI across sectors
The pattern emerging is clear. When AI lowers the cost of creation, someone downstream pays the processing cost. Australia's tribunal is the canary. Every institution that receives work from the public, applications, complaints, requests, will face this. The ones that survive will be the ones that deploy AI to manage the AI-generated flood.
The Implication
Watch for this pattern everywhere professional gatekeepers exist. University admissions offices. Patent offices. Customer service departments. Government agencies processing permits. The AI tooling that makes it easier to submit will force the deployment of AI to filter, prioritize, and process. The institutions that resist will drown in volume.
For individuals: the window where AI gives you an edge in these processes is closing fast. Once institutions deploy their own AI filters, they'll be optimized to catch AI-generated submissions. The brief advantage of using ChatGPT to draft your workplace complaint or college essay is temporary. Quality and genuine signal will matter again, just evaluated by different gatekeepers.