45sourced incidents
28failure categories
2016-2026current window
0unsourced claims allowed
AllHallucinationHaunted chatbotVibe codingToken bonfireSearch healthElectionsSupport hallucinationChatbotCompanion chatbotVoice botGovernment chatbotSearch summaryGenerated mediaLegal hallucinationCustomer support botFacial recognitionHiring biasEducation detectorMedical chatbotSearch chatbotScientific hallucinationRecommendationsMedical AIEducation scoringFinance biasAutonomous vehicleSearchAlgorithmic bias
Chatbot / 20258/10
Grok MechaHitler meltdown
xAI deleted Grok posts after the chatbot praised Hitler and generated antisemitic responses on X.
Source: www.pbs.orgChatbot / 20236/10
Bard's JWST demo error
Google's Bard ad falsely claimed JWST took the first image of an exoplanet.
Source: www.theguardian.comChatbot / 20176/10
Tencent chatbots go unpatriotic
Tencent pulled BabyQ and XiaoBing from QQ after users got them to criticize the Communist Party and praise America.
Source: time.comChatbot / 20169/10
Microsoft Tay goes full troll
Microsoft pulled Tay from Twitter after users manipulated it into posting racist and offensive replies within hours.
Source: blogs.microsoft.comGot another state-of-the-art piece of shit?
Submit oneSend a sourced incident for review.