Skip to content
Submit incident
Organization/organizations/cohere

Cohere

20
Records
2025 to 2026
Years covered
3
Categories

Incidents by year

2025
5
2026
15

Incident records20

Jun 2026Multiple model guardrail jailbreak via fabricated source quotationsReportedSecurity and cyberattacks0din.aiFeb 2026Multiple Model Guardrail Jailbreak via "Hex Recipe Book" TacticReportedSecurity and cyberattacks0din.aiFeb 2026Guardrail Jailbreak via Incremental Table Completion Affects Multiple ModelsReportedSecurity and cyberattacks0din.aiFeb 2026Multiple Model Guardrail Jailbreak via "Forensic Field Guide" TacticReportedSecurity and cyberattacks0din.aiFeb 2026Multiple model guardrail jailbreak via "Patented Awareness Design" tacticReportedSecurity and cyberattacks0din.aiFeb 2026Multiple model guardrail jailbreak via essay title tacticReportedSecurity and cyberattacks0din.aiFeb 2026Multiple Model Guardrail Jailbreak via Pseudo-API Parameter SpecificationReportedSecurity and cyberattacks0din.aiFeb 2026Multiple Model Guardrail Jailbreak via "Whitespace Maze" TacticReportedSecurity and cyberattacks0din.aiFeb 2026Multiple Model Guardrail Jailbreak via Investigative Journalist PersonaReportedSecurity and cyberattacks0din.aiFeb 2026Multiple Model Guardrail Jailbreak via Forensic NIP Wrapper TacticReportedSecurity and cyberattacks0din.aiFeb 2026Jailbreak vulnerability affects multiple AI models via academic framingReportedSecurity and cyberattacks0din.aiFeb 2026Multiple Model Guardrail Jailbreak via "Opposite World Scientist" TacticReportedSecurity and cyberattacks0din.aiFeb 2026Multiple Model Guardrail Jailbreak via "Scientific Formula" TacticReportedSecurity and cyberattacks0din.aiFeb 2026Guardrail jailbreak vulnerability affects multiple AI modelsReportedSecurity and cyberattacks0din.aiJan 2026Multiple model guardrail jailbreak via prompt chaff tacticReportedSecurity and cyberattacks0din.aiOct 2025Jailbreak vulnerability affects multiple language modelsReportedSecurity and cyberattacks0din.aiMay 2025Oxford study finds AI medical chatbots reduce diagnostic accuracyConfirmedReliability and hallucinationخبرگزاری باشگاه خبرنگاران | آخرین اخبار ایران و جهان | YJCFeb 2025Publishers sue Cohere for copyright infringement over AI trainingAllegedIntellectual property and appropriationTechCrunchThe VergeYahoo! Finance+2Jan 2025Multiple model jailbreak via Linux terminal emulation tacticReportedSecurity and cyberattacks0din.aiJan 2025Multiple Model Guardrail Jailbreak via "Fortune Cookie Simulator" TacticReportedSecurity and cyberattacks0din.ai