Public servants need right to appeal before AI flags them for poor work
A UTS report urges human review and appeals before AI can flag public servants for poor performance.
An AI dashboard can misread public servant performance. A case officer who spends extra time checking contradictory evidence might be flagged as slow.
An algorithm cannot know the context. The UTS report, released in October, calls for a fair test for monitoring and a complaints pathway through the Fair Work Commission.
Every automated alert should require a named decision-maker to inspect the evidence. That person must ask the employee for their explanation.
They must have real power to reject the recommendation. Agencies should keep a record of what data the tool analysed, what it concluded, and why the final decision followed.
Employees should know whom to contact to challenge a finding and have it reconsidered by someone independent.
- University of Technology Sydney
- Report source
- October 20
- Report release
- Named human reviewer required
- Key safeguard
- Fair Work Commission
- Escalation body
Why it mattersWithout human review, an AI system can wreck someone's career based on data it misinterprets or context it cannot see.
AustraliaThe APS is expanding algorithmic monitoring across government; these safeguards would protect Australian public servants from flawed automated decisions.
✓ Claims checked against the source and corrected before publish. checked 2 h ago
