The Consequence Library · How records are made and graded
The Australian government's automated Robodebt scheme raised roughly 470,000 to 500,000 unlawful welfare debts against more than 400,000 people over four years, was linked to at least three suicides, and produced over A$2.4 billion in refunds and compensation, a Royal Commission, and corrupt-conduct findings against former senior public servants.
Evidence caveat. Robodebt used rules-based automation, not machine learning. It is included because the authority failure it documents is the one this library exists to record, and because it is the most thoroughly investigated case in the set.
What happened
From 2015 to 2016, the Department of Human Services' Online Compliance Intervention automatically raised debts by averaging annual tax office income data against fortnightly reported income, then reversed the onus of proof onto recipients. Debt notices issued without human oversight. Internal legal advice doubting the scheme's lawfulness was suppressed, and ministers were told the scheme did not use income averaging, a statement the Royal Commission later found false. In November 2019, in Amato v Commonwealth, the Federal Court established that the averaging method did not establish a lawful debt, and the government abandoned the scheme. The 2020 class-action settlement provided A$751 million in refunds plus A$112 million in interest; a further A$548.5 million settlement was approved on 23 June 2026, bringing total redress above A$2.4 billion. The Royal Commission reported on 7 July 2023 with 57 recommendations and a sealed chapter referring individuals for civil or criminal action, describing the scheme as "a crude and cruel mechanism, neither fair nor legal". In 2026 the National Anti-Corruption Commission found two former senior public servants had engaged in serious corrupt conduct.
Where control failed
There was no legality gate. No validated legal authority was established before an automated system began issuing debts at population scale, and legal advice questioning that authority was suppressed rather than resolved. There was no pre-action control, since debts issued without any human reviewing the case, and no fail-stop despite more than three years of escalating harm signals, overturned decisions and public warnings. Oversight bodies proved ineffective throughout.
The authority question
In lawful debt recovery the authority is the legal basis for the specific claim. Here the decisions contained no case-specific evidence at all: an averaged annual number stood in for proof of a fortnightly debt, and the government ultimately conceded unlawfulness rather than defend its own records. Authority was asserted by the automation and never actually held by anyone.
What could be proven afterward
The Royal Commission reconstructed the scheme comprehensively across a 990-page report, including the suppressed legal advice and the misinformation of ministers. That reconstruction required a royal commission, years of litigation and a class action. At decision time, individual recipients faced a debt notice with no evidentiary record they could contest.
Control state, before and after
Before the consequence
No legality verification before automated debt raising. No human review of individual cases. No working stop mechanism as harm signals accumulated.
After the consequence
Scheme abandoned 2019. Class-action settlements 2020 and 2026. Royal Commission report with 57 recommendations, July 2023. NACC corrupt-conduct findings, 2026.
Sources
- Primary: Royal Commission into the Robodebt Scheme, Final Report, 7 Jul 2023 · https://robodebt.royalcommission.gov.au/publications/report
- Primary: Royal Commission media release, 7 Jul 2023 · https://robodebt.royalcommission.gov.au/news-and-media/media-releases/report-royal-commission-robodebt-scheme
- Secondary: Victoria Legal Aid, "Learning from the failures of Robodebt" · https://www.legalaid.vic.gov.au/learning-from-the-failures-of-robodebt
- Secondary: Gordon Legal, Robodebt Class Action Appeal Settlement, approved 23 Jun 2026 · https://gordonlegal.com.au/services/class-actions/robodebt-class-action-appeal/
Record history
Published 16 August 2026. Load-bearing facts re-verified against the cited sources on 16 August 2026. Corrections and material changes are appended here with their dates. To report an error in this record, write to hello@regulayer.com with the record slug and the source you believe is authoritative.
This record describes what sources establish about a consequence and the control state around it. It separates confirmed fact from source-stated cause and from architectural analysis, and it makes no claim that any control or product would have prevented the outcome. Gap codes identify a failure class, not a remedy.