AI Failure Museum
The exhibits nobody else shows.
Every AI success story has a hundred failures behind it. We catalog the postmortems — the root causes, the money lost, and the patterns that predict which AI projects will fail before they start.
Air Canada Chatbot
The chatbot made up a refund policy. The tribunal made the airline pay it.
Amazon Just Walk Out
The cashierless store that ran on 1,000 humans watching video in India
Babylon Health
From $4.2B SPAC to bankruptcy in 23 months
Boeing 737 MAX MCAS
The autonomous system that pilots couldn't override fast enough
Cigna PxDx
The system that reviewed 50 claims a second and read none of them
Dutch Toeslagenaffaire
The childcare algorithm that brought down a government
GM Cruise
$10B and eight years to learn that one incident ends the program
Google Bard Launch
One wrong sentence in a marketing GIF cost Alphabet $100 billion in a single day
Klarna AI Customer Service
The CEO who fired 700 agents, then quietly hired humans back
Mata v. Avianca
Two lawyers filed a brief with six fake cases. The judge noticed.
Microsoft Sydney
The chatbot told a New York Times reporter to leave his wife. The product shipped anyway.
Robodebt
When a government replaced caseworkers with an algorithm and got sued for A$1.8B
Sports Illustrated AI Authors
The fake bylines that cost a publisher its 70-year-old license
UnitedHealth nH Predict
An algorithm with a 90% reversal rate, used to deny care anyway
Workday AI Hiring
When the screening algorithm became the defendant
Unity's AI Monetization Debacle
How an AI-driven pricing change destroyed $7B in market cap overnight
Clearview AI's Enterprise Push
When your AI's data source becomes your biggest liability
GE Predix — The Industrial AI Platform
The $7B bet that building a platform is easier than solving a problem
Microsoft's Tay Chatbot
The 16 hours that proved AI without guardrails is a liability, not a product
Apple Card Gender Bias
When the black box gives your wife a lower credit limit — and nobody can explain why
Amazon's AI Recruiting Tool
How bias in training data turned a hiring tool into a discrimination machine
Google Health AI Diagnostics
When 94% accuracy in the lab meets 0% usability in the field
IBM Watson Health
The $4B lesson in what happens when you sell AI before you solve data
McDonald's Drive-Thru AI
Why voice AI at scale is a UX problem, not an AI problem
Target's AI Personalization Platform
The technology worked — the organization couldn't keep up
Zillow Offers
When the algorithm was right about prices and wrong about everything else
Get the next Brief
One operator. Every other Wednesday.
Plus the AI Glossary and the Failure Museum.
Real names. Real numbers. Honest analysis.
No spam. No pitch.