Skip to content
Breaches.

Breaches that exposed training data

11 breaches, most recently disclosed first. The list grows as new ones are disclosed.

OrganizationAffected
Hugging FaceAutonomous OpenAI agent breached Hugging Face and stole internal datasets and credentialsJul 172 months agoAI or modelUnknown
MetaEmployee keystroke and screen data collected to train AI was widely exposed inside MetaJun 53 months agoExposed dataUnknown
MercorAI training data contractor hit via compromised LiteLLM open source packageMar 315 months agoSupply chainUnverifiedUnknown
AnthropicDeepSeek, Moonshot and MiniMax used 24,000 fraudulent accounts to distill ClaudeFeb 237 months agoAI or modelUnknown
Scale AIScale AI left confidential AI training docs and contractor data in public Google DocsJun 24, 20251 year agoExposed dataUnverifiedUnknown
Common CrawlNearly 12,000 live API keys and passwords found in AI training datasetFeb 28, 20251 year agoExposed dataUnverifiedUnknown
SAPSAPwned: SAP AI Core flaws gave researchers access to customer cloud keys and AI artifactsJul 17, 20242 years agoExposed dataUnverifiedUnknown
ReplicateTenant isolation flaw could expose all Replicate customers' AI prompts and resultsMay 23, 20242 years agoExposed dataUnverifiedUnknown
Mistral AIEarly-access customer's employee leaked Mistral 'miqu' model weightsJan 31, 20242 years agoInsiderUnverifiedUnknown
Hugging Face1,681 exposed Hugging Face API tokens gave write access to Meta Llama and other reposDec 4, 20232 years agoExposed dataUnverifiedUnknown
RingFTC: Ring staff viewed private customer videos and used them to train algorithmsMay 31, 20233 years agoInsiderUnverifiedUnknown

New breaches by email

Wednesdays, only in weeks with newly disclosed breaches, the largest first.

Double opt-in. Unsubscribe any time.