Incident Intelligence · Real-World AI Failures

AI Incident Intelligence

Name: EAAPL Pattern Library
Creator: Enterprise AI Architecture Pattern Library
License: https://aipatterns.com.au/terms

Real public AI failures mapped to architecture patterns. Every incident surfaces the missing control, flawed assumption, or absent governance layer — and links directly to the pattern that would have prevented it.

Australian incidents cross-referenced withAEAI Incident Registry ↗— 40+ AU-specific incidents

29incidents tracked

16critical severity

39patterns implicated

29 of 29 incidents

AI Incident Intelligence

APRA Issues Supervisory Letters to 4 Regulated Entities for AI Governance Failures

Air Canada Chatbot Gave Wrong Bereavement Refund Policy

DPD Chatbot Jailbroken to Swear at Customer and Criticise DPD

Cruise Autonomous Vehicle Dragged Pedestrian 20 Feet After Collision

US Air Force AI Drone Reportedly Attacked Operator in Simulation

NatWest Cora Chatbot Gave Incorrect Guidance to Vulnerable Customers

Lawyer Cited ChatGPT-Hallucinated Non-Existent Cases in Federal Court

Allegheny Family Screening Tool Found to Flag Black Families at Higher Rates

Samsung Employees Leaked Semiconductor Trade Secrets via ChatGPT

Samsung Engineers Uploaded Proprietary Source Code and Meeting Notes to OpenAI Servers

Microsoft Bing 'Sydney' AI Made Threats and Declarations of Love

Google Bard Demo Contained Factual Error, Costing ~$100B Market Cap

Lensa AI Generated Sexualised Images From Users' Normal Selfies

OAIC Finds Bunnings Breached Privacy Act Using Facial Recognition on Customers

Clearview AI Ordered to Destroy Facial Recognition Data by Australian Regulator

Zillow Offers Algorithm Overpaid for Homes, Causing $500M+ Loss

Epic Sepsis Prediction Model Failed to Generalise, Missing Majority of Cases

LinkedIn Job Ad Algorithm Showed Tech Roles to Significantly Fewer Women

Westpac AML Monitoring Missed 23M+ Transactions — $1.3B Penalty

Australian Robodebt Automated Debt Scheme Found Unlawful, Linked to Deaths

UK Ofqual A-Level Algorithm Downgraded State School Students

UK Home Office Visa Streaming Algorithm Found Unlawfully Discriminatory

Apple Card Algorithm Allegedly Gave Women Lower Credit Limits Than Husbands

Amazon AI Recruitment Tool Systematically Discriminated Against Women

Uber ATG Autonomous Vehicle Killed Pedestrian in First Autonomous Fatality

IBM Watson for Oncology Recommended Unsafe Cancer Treatments

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

COMPAS Recidivism Algorithm Showed Racial Bias in US Courts

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