The algorithm sacked him. The Commission sent Uber the bill.

Here is a sentence that should make every business owner stop scrolling. No human was involved in the decision that ended this man’s income.

Seven years, one formula

A Canberra delivery driver had worked on Uber’s platforms since 2018. More than 18,000 rideshare trips. Close to 6,000 food deliveries. Not a single recorded complaint about a botched order. In May 2025, the system switched him off. Not a manager. Not a meeting. A formula.

His customer satisfaction score had dipped below Uber’s threshold. Of the 20 negative ratings counted against him, nine came with no explanation at all. One merchant claimed he had no delivery bag, even though his unchallenged evidence was that he always carried one. And by the time the deactivation notice arrived, his rating had actually recovered. The rolling average still dragged him under.

The Fair Work Commission found the entire process, right up to the preliminary deactivation notice, was fully automated. The human “review” that followed was an offshore team only allowed to consider five prescribed matters. Nobody in the chain had the authority to step back and ask the obvious question: does it make any sense to cut loose a near spotless seven year worker over a handful of unexplained one star ratings?

The Commission’s answer

No, it does not. The Full Bench found the deactivation unfair, ordered Uber to switch him back on, and this week put a number on it: $40,768.87 for roughly eleven months of lost income, payable within 14 days.

Sit with the human side of that for a second. Eleven months locked out of his main source of income, scrambling for extra shifts on other platforms to cover the gap, all because a calculation said so and no person could be bothered to check it. Behind every automated decision is someone with rent, a car loan and a family.

You’re not Uber. This still applies to you.

Automation is creeping into how small businesses manage people. Performance dashboards. Rostering software. AI tools that flag “underperformers”. Systems that score, rank and recommend. None of that is a problem. The problem starts the moment the tool stops informing the decision and starts making it.

Tribunals keep saying the same thing in different ways. Before someone loses their livelihood, a real human has to genuinely consider their side. A process where the outcome is locked in before anyone speaks to the person is not a process at all. It is the same principle that sank the small employer in our last case blog. No warning, no chance to respond, no fairness.

If you are using data or AI anywhere near people decisions, four rules:

• A human makes the final call, and has real authority to override the system.
• The person hears the actual reasons, specific and explained, not a score on a dashboard.
• They get a genuine chance to respond before the decision is made, not after.
• Sense check the data. An unexplained rating is an unexplained allegation, and you would never sack someone over an allegation nobody can describe.

The quiet second lesson: gig workers have real rights now

This case ran under the new employee like worker protections for digital platform workers. Gig workers can now challenge a deactivation much like employees challenge a dismissal, with reinstatement and lost pay on the table, and there is a Deactivation Code that works a lot like the Small Business Fair Dismissal Code. The gap between “employee” and “contractor” keeps narrowing. If part of your engagement model relies on “they’re contractors, so the rules don’t apply”, that assumption deserves a fresh look before someone tests it for you.

The takeaway

Technology can inform a decision about a person. It cannot own it. The moment nobody in your business can explain why someone lost their job, or override the system that decided it, you have a Fair Work problem waiting for a hearing date.

If you are bringing automation into your people management, or leaning on contractor arrangements, get the framework right before the test case is yours. Talk to HR Gurus.