naïve-labs
← AI standards
Relationship map 02 / AI infrastructure

Safety, quality& assurance.

Trustworthiness cannot be demonstrated by a single test. The map connects threat knowledge, safeguards, benchmarks, and provenance with the system in use.

Threat knowledge informs safeguards; benchmarks assess the system, while provenance standards make its content verifiable.
interactive standardsystem or outputrelationship
Safety, quality, and assurance The map connects threat knowledge, safeguards, measurement, and provenance with an AI system and its digital content. Threat knowledge Risk management, requirements, and controls Measurement and comparison OWASP Agentic Top 10key agentic risks MITRE ATLASattack techniques NIST AI RMFgeneral AI risk NIST GenAI Profileactions for GenAI ETSI EN 304 223security requirements CSA AI Controls Matrixauditable controls MLPerfperformance and energy AILuminatesafety tests C2PA Content Credentialsprovenance and edits

Light nodes open a concise description and the corresponding primary source. Actors and technical systems provide orientation.

Further developments.

OpenTelemetry GenAI Semantic Conventions Semantic conventions OpenTelemetry ↗

Status
Maintained by