57sourced incidents
31failure categories
2016-2026current window
0unsourced claims allowed
AllAI failureVibe codingHaunted chatbotGenerated mediaSearch failureAI hallucinationHallucinationToken bonfireSearch healthElectionsSupport hallucinationChatbotCompanion chatbotVoice botGovernment chatbotSearch summaryLegal hallucinationCustomer support botFacial recognitionHiring biasEducation detectorMedical chatbotSearch chatbotScientific hallucinationRecommendationsMedical AIEducation scoringFinance biasAutonomous vehicleSearchAlgorithmic bias
Generated media / 20265/10
Using ChatGPT to make text prompts is ridiculous. Look at how many times I have to...
Reddit user reports ChatGPT AI weirdness: Using ChatGPT to make text prompts is ridiculous. Look at how many times I have to...
Source: www.reddit.comGenerated media / 20247/10
Gemini historical image mess
Google paused Gemini people-image generation after historically inaccurate depictions went viral.
Source: blog.googleGenerated media / 20236/10
Sports Illustrated fake writers
Sports Illustrated removed product-review content after reporting found fake bylines and AI-looking author photos.
Source: www.pbs.orgGenerated media / 20236/10
CNET's error-prone AI finance articles
CNET corrected dozens of AI-written finance articles after basic factual and math errors surfaced.
Source: www.theverge.comGot another state-of-the-art piece of shit?
Submit oneSend a sourced incident for review.