Australian EnterpriseAI Index
AU AI Incident Database

Australian AI Incident Database

Structured, searchable database of AU AI failures, algorithmic errors, and regulatory actions. Each incident cross-links to the architectural failure patterns on aipatterns.com.au.

29

Total incidents

16

Critical

5

Sectors affected

Multi-billion

Financial impact

4Critical
1High
0Medium
0Low

Showing 5 of 29 incidents

Criticalother18 Mar 2018

Uber ATG Autonomous Vehicle Killed Pedestrian in First Autonomous Fatality

Uber Advanced Technologies Group

Details

An Uber self-driving test vehicle struck and killed Elaine Herzberg in Tempe, Arizona. The system detected the pedestrian but misclassified her, and the automatic emergency braking was disabled during autonomous operation.

NHTSA Automated Vehicle PolicyArizona AV Testing RegulationsNTSB Investigation Standardshuman override pattern ai confidence threshold routing
Criticalhealthcare25 July 2017

IBM Watson for Oncology Recommended Unsafe Cancer Treatments

IBM Watson Health

Details

Internal IBM documents leaked to STAT News revealed that IBM Watson for Oncology had recommended treatments that oncologists described as 'unsafe and incorrect'. The system generated advice based on hypothetical training cases rather than real patient data.

FDA SaMD Guidance21 CFR Part 820 (QSR)HIPAAai training data governance ai performance benchmarking
Criticalgovernment1 July 2016

Robodebt: Automated Income-Averaging Issued 381,000 Unlawful Debt Notices

Services Australia (Department of Human Services)

Details

The Australian Government's automated welfare debt recovery scheme used income-averaging to calculate alleged debts owed by welfare recipients, comparing Centrelink payment data against ATO tax records without human verification. The algorithm generated 381,000 debt notices over three years.

Privacy ActSocial Security ActAdministrative Decisions (Judicial Review) Acthuman approval gateway ai audit trail
Criticalgovernment23 May 2016

COMPAS Recidivism Algorithm Showed Racial Bias in US Courts

Northpointe (Equivant)

Details

ProPublica's investigation found that Black defendants were nearly twice as likely to be falsely flagged as future criminals compared to white defendants. The tool was being used by judges to inform bail, sentencing, and parole decisions.

US Civil Rights ActEqual Protection ClauseEU AI Act (Prohibited/High-Risk)model bias detection responsible ai framework
Highother17 June 2014

Meta Manipulated 700,000 Users' News Feeds Without Consent in Emotion Study

Meta (Facebook)

Details

Facebook conducted a large-scale psychological experiment on approximately 689,000 users, systematically altering their News Feed algorithms to test whether emotions were contagious online. The study was published without users ever being informed they were research subjects.

FTC Act Section 5Common Rule (45 CFR 46)GDPR (analogous)ai ethics review board privacy by design for ai data
PreviousPage 2 of 2
Admin-curated database. Incidents are sourced from OAIC, ASIC, APRA, AUSTRAC, TGA, Federal Court records, and verified reporting. Contact us with sources for consideration.Pattern cross-links reference architectural failure categories on aipatterns.com.au. Not legal advice.