AI TRiSM

AI Trust, Risk and Security Management (AI TRiSM) is a Gartner coinage that bundles together the AI-adjacent control stack: AI model governance (lifecycle controls, inventory, approvals), AI-SPM (security posture of models and model-data), AI runtime enforcement (prompt filtering, output moderation, abuse detection), explainability tooling, data protection (training-data provenance, inference-time personal-data handling), and bias / fairness controls. The framework is more useful as a shopping list than as an architecture — organisations typically assemble pieces from different vendors rather than buying 'AI TRiSM' as a single product.

For regulators the four pillars line up with specific obligations: AI governance maps to the EU AI Act Articles 9–17 and the DPDP Section 10 SDF duties; AI-SPM evidence supports GDPR Article 35 DPIAs; explainability is the technical satisfaction of the Article 22 right to meaningful information about the logic involved; and bias / fairness controls support both EU AI Act Article 10 data-governance and non-discrimination duties across jurisdictions. Mature programmes bind AI TRiSM into the privacy and security stacks rather than treating it as a parallel vertical.