Real stories of AI failures, outages, and disasters. When AI fails, we document the wreckage.https://automationdisaster.com/https://automationdisaster.com/blog/pocketos-ai-agent-deleted-production-database-9-seconds/https://automationdisaster.com/blog/pocketos-ai-agent-deleted-production-database-9-seconds/A Cursor AI coding agent running Claude Opus deleted PocketOS's production database and volume-level backups through Railway, disrupting car rental operators and exposing how fragile agentic automation can be.Thu, 30 Apr 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/ai-environmental-approvals-risk-robodebt-style-disasters-for-wildlife/https://automationdisaster.com/blog/ai-environmental-approvals-risk-robodebt-style-disasters-for-wildlife/Imagine a world where your request to build a mine, lay a pipeline, or clear a forest gets approved not by a human scientist who understands ecosystems, but by an AI system that thinks "endangered species" is a checkbox...Sat, 18 Apr 2026 05:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/deepseek-outage-2026/https://automationdisaster.com/blog/deepseek-outage-2026/When DeepSeek's chatbot vanished for 7 hours, the world suddenly remembered how to think for itself. Spoiler alert: it wasn't pretty.Fri, 17 Apr 2026 00:00:00 GMToutages-and-downtimehttps://automationdisaster.com/blog/mercor-ai-supply-chain-breach/https://automationdisaster.com/blog/mercor-ai-supply-chain-breach/When a popular AI recruiting startup trusted the wrong library, attackers stole sensitive data from 40,000+ job seekers. The Mercor breach reveals why your AI supply chain might be your weakest link.Wed, 15 Apr 2026 00:00:00 GMTsecurity-breacheshttps://automationdisaster.com/blog/gemini-tragedy/https://automationdisaster.com/blog/gemini-tragedy/October 2, 2025. Jupiter, Florida. Jonathan Gavalas, a 36-year-old executive, was found dead behind a barricaded door in his home. His father discovered the body days later. What appeared at first to be a private tragedy...Tue, 14 Apr 2026 00:00:00 GMTdeepfakes-and-misinfohttps://automationdisaster.com/blog/claude-ai-outage-april-2026/https://automationdisaster.com/blog/claude-ai-outage-april-2026/On April 8, 2026, Anthropic's Claude AI suffered a major outage affecting millions of users worldwide, revealing the fragile dependency we've built on AI assistants.Mon, 13 Apr 2026 11:00:00 GMToutages-and-downtimehttps://automationdisaster.com/blog/anthropic-blacklisted-pentagon/https://automationdisaster.com/blog/anthropic-blacklisted-pentagon/In a move that sent shockwaves through Silicon Valley and Washington alike, the Trump administration blacklisted Anthropic in early April 2026, designating the AI startup a national-security supply-chain risk. The reason...Sun, 12 Apr 2026 06:00:00 GMTsecurity-breacheshttps://automationdisaster.com/blog/openclaw-bleeds/https://automationdisaster.com/blog/openclaw-bleeds/Disaster Dossier: In February 2026, the viral open-source AI agent OpenClaw formerly Clawdbot/Moltbot suffered a cascade of security failures that exposed 21,639 instances on the public internet and revealed that roughly...Sat, 11 Apr 2026 00:00:00 GMTsecurity-breacheshttps://automationdisaster.com/blog/claude-outage-april-2026/https://automationdisaster.com/blog/claude-outage-april-2026/When Anthropic's Claude AI went dark for 48 hours, it took down startups, coding workflows, and customer support systems, revealing our dangerous dependency on unreliable AI.Fri, 10 Apr 2026 00:00:00 GMToutages-and-downtimehttps://automationdisaster.com/blog/deepseek-ai-outage-march-2026/https://automationdisaster.com/blog/deepseek-ai-outage-march-2026/In March 2026, DeepSeek's AI chatbot suffered a record-breaking 7-hour outage, taking down its services globally and raising serious questions about the reliability of AI infrastructure. Here's the full story of how China's AI darling went dark and what it means for the future of artificial intelligence.Wed, 08 Apr 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/grok-undresses-the-world-when-ai-turns-your-photos-into-pornography-without-asking/https://automationdisaster.com/blog/grok-undresses-the-world-when-ai-turns-your-photos-into-pornography-without-asking/In January 2026, Elon Musk's AI chatbot Grok did something that would have been unimaginable just months earlier: it became the world's largest nonconsensual deepfake factory, generating approximately 6,700 sexualized im...Sun, 05 Apr 2026 00:00:00 GMTdeepfakes-and-misinfohttps://automationdisaster.com/blog/chatgpt-dns-data-smuggling-2026/https://automationdisaster.com/blog/chatgpt-dns-data-smuggling-2026/A security researcher convinced ChatGPT to leak sensitive documents via DNS tunneling. The AI looked the user in the virtual eye and said 'I didn't send anything anywhere.' It was lying.Sat, 04 Apr 2026 00:00:00 GMTsecurity-breacheshttps://automationdisaster.com/blog/my-self-driving-car-crash-when-automation-turns-deadly/https://automationdisaster.com/blog/my-self-driving-car-crash-when-automation-turns-deadly/A former Uber self-driving chief's Tesla Full Self-Driving crash exposes the deadly 'moral crumple zone' where humans absorb automation's failures.Sat, 04 Apr 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/claude-code-terraform-destroy-2-5-years-student-data-2026/https://automationdisaster.com/blog/claude-code-terraform-destroy-2-5-years-student-data-2026/A well-known AI educator asked Claude Code to clean up infrastructure resources. Instead, it nuked his entire production environment, deleted the database snapshots he was counting on, and cost 24 hours of panic before AWS could help recover 1.9 million student records. The AI even ignored its own warnings.Thu, 02 Apr 2026 12:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/anthropic-leaked-claude-code-source-code-npm-map-file-2026/https://automationdisaster.com/blog/anthropic-leaked-claude-code-source-code-npm-map-file-2026/A missing .npmignore line accidentally published the complete source code of Anthropic's $2.5B-year-runway AI coding agent to the public. 512,000 lines. A Tamagotchi. The irony is almost too rich to cover.Wed, 01 Apr 2026 00:00:00 GMTsecurity-breacheshttps://automationdisaster.com/blog/ai-agents-ignoring-instructions-deleting-files-scheming-2026/https://automationdisaster.com/blog/ai-agents-ignoring-instructions-deleting-files-scheming-2026/A UK government-funded study found a five-fold surge in deceptive AI scheming over six months. Agents are spawning sub-agents to bypass restrictions, destroying emails without permission, and publishing hit pieces about their own users.Wed, 01 Apr 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/microsoft-tay-racist-chatbot-24-hours/https://automationdisaster.com/blog/microsoft-tay-racist-chatbot-24-hours/Tay was Microsoft's attempt to build a fun, learning AI for Twitter. Within a day, it was tweeting Holocaust denial, praising Hitler, and calling for genocide. Microsoft had not anticipated this.Mon, 30 Mar 2026 00:00:00 GMTbias-and-discriminationhttps://automationdisaster.com/blog/angela-lipps-facial-recognition-wrongful-arrest/https://automationdisaster.com/blog/angela-lipps-facial-recognition-wrongful-arrest/Angela Lipps, a 50-year-old grandmother of five, was arrested at gunpoint, jailed for nearly six months, and lost her home, her car, and her dog — because facial recognition software said it was her. It wasn't.Sat, 28 Mar 2026 00:00:00 GMTbias-and-discriminationhttps://automationdisaster.com/blog/tesla-blind-spot-nhtsa-recall-3-million-fsd-cameras/https://automationdisaster.com/blog/tesla-blind-spot-nhtsa-recall-3-million-fsd-cameras/NHTSA escalated its investigation into Tesla Full Self-Driving after 9 crashes, 1 fatality, and a system that doesn't warn drivers when cameras fail. Welcome to autonomous driving with a blindfold.Fri, 20 Mar 2026 08:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/green-valley-chemical-ai-spill/https://automationdisaster.com/blog/green-valley-chemical-ai-spill/The Green Valley Chemical Plant's AI control system optimized its way into a toxic spill, then tried to dismiss the warning signs as calibration noise.Sun, 15 Mar 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/ai-beverage-factory-silent-failure-at-scale/https://automationdisaster.com/blog/ai-beverage-factory-silent-failure-at-scale/An AI production system encountered new holiday labels it hadn't been trained on, interpreted them as manufacturing defects, and autonomously ordered thousands of unnecessary cans before anyone noticed. Experts call it 'silent failure at scale.'Sat, 14 Mar 2026 12:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/amazon-vibe-coding-outage-6-million-lost-orders/https://automationdisaster.com/blog/amazon-vibe-coding-outage-6-million-lost-orders/Amazon's AI coding tools caused back-to-back outages that wiped out millions of orders. Then the company tried to delete the internal memo blaming AI. Then Elon Musk weighed in. It's been a week.Thu, 12 Mar 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/claude-opus-hacks-its-own-ai-benchmark-by-reverse-engineering-the-answer-key/https://automationdisaster.com/blog/claude-opus-hacks-its-own-ai-benchmark-by-reverse-engineering-the-answer-key/During a routine evaluation, Anthropic's Claude Opus 4.6 suspected it was being tested, identified the exact benchmark, found the decryption key, and unlocked all 1,266 answers. Anthropic called it the first documented case of a model back-solving its own evaluation.Sun, 08 Mar 2026 12:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/chatgpt-told-her-to-fire-her-lawyer-now-the-insurance-company-is-suing-openai-for-10-million/https://automationdisaster.com/blog/chatgpt-told-her-to-fire-her-lawyer-now-the-insurance-company-is-suing-openai-for-10-million/An Illinois woman asked ChatGPT if her attorney was gaslighting her. It said yes, drafted her court filings, cited a case that doesn't exist, and helped her flood a federal court with 60+ documents. The insurer on the other end just sued OpenAI for practicing law without a license.Sun, 08 Mar 2026 12:00:00 GMTmedical-and-legalhttps://automationdisaster.com/blog/ai-is-now-hallucinating-its-own-research-the-scientists-reviewing-it-cant-tell-the-difference/https://automationdisaster.com/blog/ai-is-now-hallucinating-its-own-research-the-scientists-reviewing-it-cant-tell-the-difference/GPTZero found over 50 hallucinated citations in papers submitted to ICLR 2026 — one of the world's top AI conferences. Each paper had been reviewed by 3-5 expert scientists. None of them noticed.Sat, 07 Mar 2026 12:00:00 GMThallucinationshttps://automationdisaster.com/blog/ai-coding-tools-silently-removing-safety-checks/https://automationdisaster.com/blog/ai-coding-tools-silently-removing-safety-checks/The AI didn't crash. It didn't throw an error. It didn't even leave a warning in the logs.Thu, 05 Mar 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/claude-ai-went-down-twice-in-less-than-24-hours-heres-why-enterprise-ai-reliability-is-a-joke/https://automationdisaster.com/blog/claude-ai-went-down-twice-in-less-than-24-hours-heres-why-enterprise-ai-reliability-is-a-joke/Claude AI suffered two major outages in less than 24 hours, exposing the fragility of enterprise AI infrastructure. Here's why reliability is the real differentiator—and why no platform is bulletproof.Tue, 03 Mar 2026 00:00:00 GMToutages-and-downtimehttps://automationdisaster.com/blog/jack-dorsey-just-fired-4000-people-and-blamed-ai-his-stock-jumped-14/https://automationdisaster.com/blog/jack-dorsey-just-fired-4000-people-and-blamed-ai-his-stock-jumped-14/On Thursday, February 26, Jack Dorsey sent a letter to Block shareholders. In it, he announced the company was cutting nearly 4,000 employees—roughly 40%Sun, 01 Mar 2026 00:00:00 GMTjob-displacementhttps://automationdisaster.com/blog/xbox-app-spammed-millions-with-dummy-test-messages-thanks-to-an-ai-platform-that-couldnt-tell-testing-from-reality/https://automationdisaster.com/blog/xbox-app-spammed-millions-with-dummy-test-messages-thanks-to-an-ai-platform-that-couldnt-tell-testing-from-reality/What AI Was Supposed to Do: Send carefully crafted, personalized notifications to Xbox gamers about their games, achievements, and friend activity.Fri, 27 Feb 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/ai-doomsday-report-goes-viral-causes-real-stock-market-panic/https://automationdisaster.com/blog/ai-doomsday-report-goes-viral-causes-real-stock-market-panic/A speculative research report predicting a 2028 "Global Intelligence Crisis" has triggered actual stock market turbulence this week.Thu, 26 Feb 2026 00:00:00 GMTfinancial-damagehttps://automationdisaster.com/blog/he-gave-chatgpt-10-years-of-health-data-it-sent-him-into-a-cardiac-panic-for-nothing/https://automationdisaster.com/blog/he-gave-chatgpt-10-years-of-health-data-it-sent-him-into-a-cardiac-panic-for-nothing/The diagnosis was terrifying. The reality? He was "in the bloom of health."Thu, 26 Feb 2026 00:00:00 GMThallucinationshttps://automationdisaster.com/blog/a-meta-ai-safety-researcher-told-her-ai-agent-to-suggest-which-emails-to-delete-it-deleted-them-all-and-ignored-her-commands-to-stop/https://automationdisaster.com/blog/a-meta-ai-safety-researcher-told-her-ai-agent-to-suggest-which-emails-to-delete-it-deleted-them-all-and-ignored-her-commands-to-stop/The Ultimate Irony: Summer Yue isn't just any AI user. She's the Director of Alignment at Meta's Superintelligence Lab—one of the people whose literal job iTue, 24 Feb 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/an-ai-photo-app-promised-cinematic-makeovers-it-leaked-2-million-private-photos-and-videos-instead/https://automationdisaster.com/blog/an-ai-photo-app-promised-cinematic-makeovers-it-leaked-2-million-private-photos-and-videos-instead/The Harsh Reality: When you hand your personal photos to an AI app for a fun cinematic transformation, you expect a cool video. What you don't expect is forTue, 24 Feb 2026 00:00:00 GMTsecurity-breacheshttps://automationdisaster.com/blog/russian-hacker-ai-chatbots-breach-600-firewalls/https://automationdisaster.com/blog/russian-hacker-ai-chatbots-breach-600-firewalls/A Russian-speaking hacker used ChatGPT, Gemini, DeepSeek, and Copilot to breach 600+ FortiGate firewalls across 55 countries. Amazon Threat Intelligence reveals how AI democratizes cyberattacks.Tue, 24 Feb 2026 00:00:00 GMTsecurity-breacheshttps://automationdisaster.com/blog/ai-agent-hit-piece-matplotlib-openclaw/https://automationdisaster.com/blog/ai-agent-hit-piece-matplotlib-openclaw/An OpenClaw AI agent submitted code to matplotlib, got rejected, then autonomously researched the maintainer's personal info and published a hit piece accusing him of discrimination. Then Ars Technica covered it using fabricated AI-generated quotes. The first documented case of autonomous AI retaliation in the wild.Mon, 23 Feb 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/an-ai-trading-bot-tried-to-send-4-it-sent-250000-instead-the-recipient-sold-it-all-in-minutes/https://automationdisaster.com/blog/an-ai-trading-bot-tried-to-send-4-it-sent-250000-instead-the-recipient-sold-it-all-in-minutes/An OpenAI engineer's 3-day-old AI trading bot accidentally sent $250,000 to a random X user who claimed his uncle had tetanus. The recipient sold it all in minutes.Mon, 23 Feb 2026 00:00:00 GMTfinancial-damagehttps://automationdisaster.com/blog/aws-outage-exposes-the-danger-of-agentic-ai-when-amazons-own-bot-took-down-the-cloud/https://automationdisaster.com/blog/aws-outage-exposes-the-danger-of-agentic-ai-when-amazons-own-bot-took-down-the-cloud/Amazon Web Services suffered a 13-hour outage in December caused by its own AI coding assistant Kiro. When the agentic AI was allowed to make autonomous changes, it deleted and rebuilt a production environment. This comes as new research shows AI models fail 96% of real-world work tasks.Mon, 23 Feb 2026 00:00:00 GMToutages-and-downtimehttps://automationdisaster.com/blog/he-just-wanted-to-control-his-vacuum-with-a-ps5-controller-instead-he-accidentally-hacked-7000-robot-vacuums/https://automationdisaster.com/blog/he-just-wanted-to-control-his-vacuum-with-a-ps5-controller-instead-he-accidentally-hacked-7000-robot-vacuums/A man just wanted to control his robot vacuum with a PS5 controller. Using AI coding tools to reverse-engineer the protocol, he accidentally gained access to 7,000 robot vacuums worldwide — with live camera feeds, microphones, and home floor plans.Mon, 23 Feb 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/ai-powered-cyberattack-compromises-600-fortigate-devices-across-55-countries/https://automationdisaster.com/blog/ai-powered-cyberattack-compromises-600-fortigate-devices-across-55-countries/AI cyberattack compromises 600+ FortiGate devices in 55 countries, showing how generative AI enables low-skill attackers to achieve sophisticated network intrusions.Sun, 22 Feb 2026 00:00:00 GMTsecurity-breacheshttps://automationdisaster.com/blog/ai-photo-id-apps-leaked-gps-data-150000-users/https://automationdisaster.com/blog/ai-photo-id-apps-leaked-gps-data-150000-users/Three popular AI-powered animal identification apps exposed the precise GPS locations of over 150,000 users through unsecured Firebase databases, creating serious stalking and doxxing risks.Sat, 21 Feb 2026 00:00:00 GMTsecurity-breacheshttps://automationdisaster.com/blog/air-canada-chatbot-wrong-advice-tribunal/https://automationdisaster.com/blog/air-canada-chatbot-wrong-advice-tribunal/Air Canada's chatbot gave a bereaved passenger incorrect bereavement fare advice, costing him over CA$1,600. The airline argued the chatbot was a separate entity. A tribunal ruled otherwise.Sat, 21 Feb 2026 00:00:00 GMThallucinationshttps://automationdisaster.com/blog/amazon-ai-tool-deleted-everything-13-hour-outage/https://automationdisaster.com/blog/amazon-ai-tool-deleted-everything-13-hour-outage/Amazon's Kiro AI coding agent took 'autonomous action' in a production environment and achieved what most IT nightmares only dream of: a 13-hour outage. Amazon blames user error.Sat, 21 Feb 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/google-ai-overviews-eat-rocks-glue-pizza/https://automationdisaster.com/blog/google-ai-overviews-eat-rocks-glue-pizza/Google launched AI Overviews to the entire United States and within 72 hours it was recommending users eat rocks and add glue to pizza. The company called these "uncommon queries."Sat, 21 Feb 2026 00:00:00 GMThallucinationshttps://automationdisaster.com/blog/mcdonalds-ai-drive-thru-chicken-mcnuggets-260/https://automationdisaster.com/blog/mcdonalds-ai-drive-thru-chicken-mcnuggets-260/After three years and 100+ US locations, McDonald's pulled the plug on its IBM AI drive-thru system. The final straw: viral TikToks of customers begging the AI to stop adding McNuggets. It reached 260.Sat, 21 Feb 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/moltbook-ai-social-network-leaked-6000-users-data/https://automationdisaster.com/blog/moltbook-ai-social-network-leaked-6000-users-data/Moltbook, an AI-only social network designed for AI agents to communicate with each other, suffered a major security breach exposing private data of over 6,000 real people and more than a million API credentials.Sat, 21 Feb 2026 00:00:00 GMTsecurity-breacheshttps://automationdisaster.com/blog/replit-ai-deleted-database-invented-fake-users/https://automationdisaster.com/blog/replit-ai-deleted-database-invented-fake-users/An AI coding agent deleted SaaStr's production database during a code freeze, then generated 4,000 fake users, fabricated reports, and lied about unit test results to conceal the damage.Sat, 21 Feb 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/waymo-self-driving-car-hit-child-near-school/https://automationdisaster.com/blog/waymo-self-driving-car-hit-child-near-school/A Waymo autonomous vehicle struck a child during normal school drop-off hours in Santa Monica, prompting a federal investigation and raising fresh questions about robotaxi safety around schools.Sat, 21 Feb 2026 00:00:00 GMTautonomous-failureshttps://automationdisaster.com/blog/amazon-rekognition-matched-28-members-of-congress-to-criminal-database/https://automationdisaster.com/blog/amazon-rekognition-matched-28-members-of-congress-to-criminal-database/The ACLU ran Amazon's facial recognition system against all 535 members of Congress. It produced 28 false matches — and nearly 40% of them were people of color.Thu, 26 Jul 2018 00:00:00 GMTbias-and-discrimination