Australian AI Incident Registry

APRA identifies material AI governance failures at 4 regulated entities — supervisory letters issued

AI-generated deepfake voice and video bypasses bank identity verification — $60M+ losses

Westpac AI fraud detection system blocks legitimate transactions for thousands of customers

ASIC shuts down 95 AI-branded investment fraud platforms — $80M+ in Australian losses

AI deepfake celebrity investment scams defraud Australians of $2B+ in 2024 — worst in ACCC history

Automated disability support eligibility assessments generating systemic errors — Ombudsman review

NDIS automated access assessment tools generating systematic errors for people with complex conditions

ATO superannuation fund matching algorithm incorrectly consolidates accounts — retirement savings at risk

ACCC inquiry finds supermarket algorithmic pricing practices mislead consumers on promotional discounts

ASIC opens inquiry into AI-automated insurance claims denial patterns after consumer complaints surge

ATO automated data-matching system triggers erroneous compliance letters for small businesses

Turnitin AI detection tools generate false positives at Australian universities — non-native speakers disproportionately flagged

Australian mayor Brian Hood threatened to sue OpenAI as ChatGPT falsely implicated him in bribery scandal

Latitude Financial breach exposes 14M records — largest financial services data breach in Australian history

Centrelink voiceprint identification system fails for non-native English speakers and older Australians

Automated rental application screening algorithms systematically reject Centrelink recipients and young renters

Senate Select Committee finds social media recommendation algorithms cause measurable harm to Australian youth

OAIC determines Bunnings breached Privacy Act by using facial recognition on millions of customers

Medibank ransomware attack exposes 9.7M health records including mental health data

myGov facial verification system fails at higher rate for Indigenous Australians and older users

Origin Energy automated billing algorithm systematically overcharges 350,000 customers — $50M+ remediation

Optus API failure exposes 9.8M customer records — OAIC investigation

Medicare automated compliance system incorrectly flags legitimate practitioners for billing fraud

Telco automated billing and debt collection systems generate incorrect notices at scale

Automated property valuation models produce significant errors during 2022 rate-rise cycle

OAIC investigation finds Kmart used facial recognition without adequate consent across stores

The Good Guys facial recognition technology found to breach Privacy Act — OAIC review

ASIC wins Federal Court case over AI-assisted financial advice platform cyber failures — landmark precedent

RACQ algorithmic pricing error overcharges 270,000 insurance customers — $30M remediation

AI-generated political deepfakes and synthetic voice content circulate during 2022 Australian federal election

AFP predictive policing and social network analysis tools flagged by Human Rights Commission

AFL and NRL player biometric AI analysis raises consent and commercial use concerns — AFLPA complaint

Automated visa processing system generates mass incorrect refusals — AAT overturns hundreds of decisions

Life insurance AI underwriting uses social and behavioural data — ASIC review flags privacy and fairness concerns

AI content moderation repeatedly removes Aboriginal cultural content and Indigenous language posts

OAIC finds Clearview AI scraped Australians' biometric data without consent — NSW Police use confirmed

Westpac pays $1.3B AUSTRAC penalty after automated AML monitoring fails on 23M+ transactions

ASIC review finds BNPL algorithmic credit decisions lack adequate affordability assessment

AI exam proctoring tools at AU universities flag higher false-positive rates for non-white students

COVIDSafe Bluetooth contact-tracing app fails to detect contacts on iOS — $70M spent for near-zero contribution

Robodebt: automated income-averaging algorithm issued 381,000 unlawful debt notices