Organization/organizations/truthful-ai
Truthful AI
Incidents by year
20251
20262
Incident records3
Apr 2026LLMs transmit behavioral traits via hidden signals in training dataNaturenature.comIFLScience+3Jan 2026AI Models Show Emergent Misalignment, Generate Violent OutputsLa VanguardiaLa Voz de GaliciaJul 2025Anthropic study reveals AI models inherit risky behaviors through subliminal learningThe Times of India