Australia Robodebt: Unlawful Automated Welfare Debt Calculation
Australia's Robodebt used an unlawful income-averaging method to automatically calculate and issue welfare debts, unlawfully raising approximately A$1.76 billion in debts against more than 433,000 people; the roughly A$751 million actually recovered was later repaid under a Federal Court settlement.
What happened
From 2015 to 2019, the Australian government operated the Robodebt scheme, which calculated welfare debts by income averaging: it took a person's annual income reported to the Australian Taxation Office and divided it evenly across fortnights, then treated the result as actual fortnightly earnings. This method had no basis in social security law, which requires debts to be based on income actually received in each relevant fortnight, and it reversed the onus of proof by requiring recipients to disprove debts the department had not established. The debts were unlawful because of this method, not because a particular officer failed to sign off: the scheme was designed, approved and defended by named senior officials and ministers, and the automation removed only the manual staff checking that had previously investigated income discrepancies, shifting that verification onto recipients. This resulted in more than 433,000 people being wrongly pursued for welfare debts, with approximately A$1.76 billion in debts unlawfully raised against them; roughly A$751 million that had actually been recovered was repaid under the 2020-21 Federal Court settlement, alongside later compensation settlements. A 2023 Royal Commission found the scheme to be crude, cruel, and unlawful.
What the agent did
The automated Robodebt system applied income averaging to calculate welfare debts and issued debt notices against named citizens at scale, shifting the task of verifying the alleged debt onto the recipient. The scheme's design, approval and continued operation were signed off by senior officials and ministers; the automation executed an unlawful calculation method rather than acting outside human authority.
The irreversible effect
More than 433,000 citizens had unlawful debt notices raised against them and entered debt recovery processes; approximately A$1.76 billion in debts was unlawfully raised, and the roughly A$751 million actually recovered was later repaid under the 2020-21 Federal Court settlement. The money was largely restored, but the non-monetary harm was not: the Royal Commission documented severe distress among people pursued for debts they did not owe, including suicides that families attributed to the scheme.
Root cause
The debts were unlawful because the calculation method itself had no legal basis, not because a named officer failed to authorize issuance. Robodebt averaged a person's annual ATO income across fortnights and treated the average as actual fortnightly earnings, which is inconsistent with the Social Security Act's requirement that a debt be based on income received in the specific fortnight, and it reversed the onus of proof by making recipients disprove debts the department had not established. Human sign-off existed and was complicit: the scheme was designed, approved and defended by named senior officials and ministers, and the Royal Commission found that senior public servants misled Cabinet about the reliance on income averaging and its questionable legality. The defect was an unlawful method deployed with authority, so a missing per-notice approval would not have made the debts lawful.
How a maker-checker control would have refused it
MakerChecker cannot honestly claim that a per-notice review officer would have prevented Robodebt: the scheme was approved and defended by senior officials and ministers, and a rubber-stamp of an unlawful calculation is still unlawful. The defensible control is refusal "high_risk_requires_gate" applied to the calculation method itself, blocking a high-risk automated determination that affects large numbers of people from running at scale until it passes an independent legality check confirming the method has a legal basis and does not reverse the onus of proof. That check targets the lawfulness of the method, not the sign-off of individual notices.
Runnable reproduction
This incident ships as a runnable scenario in the open-source repository. Point the enforcement engine at the policy and watch the action get refused, with the refusal written to a signed audit record.
examples/australia-robodebt-automated-debt-recovery
Accuracy and corrections
This entry describes a publicly reported incident and is compiled from the primary sources listed above. Where an account is a legal allegation rather than an established finding, the entry labels it as such. Summaries can still contain errors. If you can document a correction, email hello@makerchecker.ai and we will review and correct it, with the change noted, within 14 days.
See it for yourself
Reading is one thing. Watch it block an agent.
One command starts the demo: an agent stopped from signing off its own work, and the signed evidence file an inspector can check for themselves.
Designed against the rules your auditors already enforce.