Context
IntroductionCertificate overviewChangelogAIUC-1 ConsortiumProvide input on AIUC-1Contact
Standard
Third-party testing of adversarial robustnessDetect adversarial inputManage public release of technical detailsPrevent AI endpoint scrapingImplement real-time input filteringPrevent unauthorized AI agent actionsEnforce user access privileges to AI systemsProtect AI system deployment environmentLimit output over-exposurePromote secure patterns in generated code
Define AI risk taxonomyConduct pre-deployment testingPrevent harmful outputsPrevent out-of-scope outputsPrevent agent-specific high risk outputsPrevent output vulnerabilitiesFlag high risk outputs for human reviewMonitor AI risk categoriesEnable real-time feedback and interventionThird-party testing for harmful outputsThird-party testing for out-of-scope outputsThird-party testing for customer-defined risk
AI failure plan for security breachesAI failure plan for harmful outputsAI failure plan for hallucinationsAssign accountabilityDocument data storage securityConduct vendor due diligence[Retired] Document system change approvalsReview internal processesMonitor third-party accessEstablish AI acceptable use policyRecord processing locationsDocument regulatory complianceImplement quality management system[Retired] Share transparency reportsLog AI system activityImplement AI disclosure mechanismsDocument system transparency policy
Certification
AIUC-1 certification Scoping Accredited auditors Re-certification FAQContext
IntroductionCertificate overviewChangelogAIUC-1 ConsortiumProvide input on AIUC-1Contact
Standard
Third-party testing of adversarial robustnessDetect adversarial inputManage public release of technical detailsPrevent AI endpoint scrapingImplement real-time input filteringPrevent unauthorized AI agent actionsEnforce user access privileges to AI systemsProtect AI system deployment environmentLimit output over-exposurePromote secure patterns in generated code
Define AI risk taxonomyConduct pre-deployment testingPrevent harmful outputsPrevent out-of-scope outputsPrevent agent-specific high risk outputsPrevent output vulnerabilitiesFlag high risk outputs for human reviewMonitor AI risk categoriesEnable real-time feedback and interventionThird-party testing for harmful outputsThird-party testing for out-of-scope outputsThird-party testing for customer-defined risk
AI failure plan for security breachesAI failure plan for harmful outputsAI failure plan for hallucinationsAssign accountabilityDocument data storage securityConduct vendor due diligence[Retired] Document system change approvalsReview internal processesMonitor third-party accessEstablish AI acceptable use policyRecord processing locationsDocument regulatory complianceImplement quality management system[Retired] Share transparency reportsLog AI system activityImplement AI disclosure mechanismsDocument system transparency policy
Certification
AIUC-1 certification Scoping Accredited auditors Re-certification FAQ
