Evidence Library
Evidence Library
Evidence-first AI assurance resources for mapping systems, risks, controls, evidence, human review, gaps, and domain-specific documentation.
Library Index
Evidence HubGeneral flagship operational resource
AI Assurance Evidence Review Kit v1.0
A seven-worksheet Excel workflow for profiling an AI system, mapping risks to controls and evidence, documenting human review, assessing evidence quality, recording gaps, and preparing a restrained management summary.
The Kit uses a coherent synthetic AML alert-prioritization example to demonstrate a general-purpose method. Evidence-quality scoring supports triage and does not establish control effectiveness, compliance, evidence sufficiency, or audit acceptance.
Start here when you need
- A bounded AI system profile
- A risk-control-evidence relationship record
- Decision-level human oversight evidence
- Transparent evidence-quality triage
- Gap, owner, target-date, and management follow-up
Specialized AML public research prototype
AML AI Assurance Evidence Pack v0.2
A separate AML-specific documentation prototype for system context, monitoring method context, alert review and escalation, human-oversight references, model and data events, evidence sources, limitations, and consistency checks.
The AML Pack retains its own publication status and limitations. It does not establish compliance, control effectiveness, model or dataset validity, evidence sufficiency, effective human oversight, regulatory acceptance, or production readiness.
Domain resources
Focused Evidence Topics
AML AI Oversight
Reviewer authority, alert escalation, rationale, overrides, automation-bias controls, and evidence traces for financial-crime workflows.
Model Risk Support
Evidence topics for model governance, monitoring, change, limitations, validation inputs, and AI model-risk review.
Vendor AI Risk Evidence
Third-party documentation, version changes, limitations, testing evidence, contract dependencies, and oversight records.
Supporting resources
Templates, Crosswalks, and Research
Templates
Supporting governance and evidence templates may complement the flagship Kit, but the Templates page is not the Kit’s canonical landing page.
AI Risk
Risk-domain context and related research that can inform organization-specific risk identification and evidence planning.
Vendor AI Risk
Broader third-party AI risk context supporting evidence requests and vendor-governance decisions.
LLM / RAG Risk
Risk and control context for retrieval-augmented and language-model systems.
Editorial Standards
Publication, sourcing, evidence, and claim-control principles for InfoSecured.ai research materials.
GridLock GRC
A public proof-of-work concept for linking risks, controls, evidence objects, owners, oversight records, and review decisions.
Use boundaries
Treat Each Artifact According to Its Own Status
The general Kit and the AML Pack are distinct artifacts with different scope and publication language. Do not transfer the AML Pack’s OSG-009 status, professional-review status, license status, artifact identifiers, or verification claims to the general Kit.
Public materials are educational and research-oriented resources. Validate applicability, configuration, security, decision rights, approvals, monitoring, and evidence retention for the organization and use case.
Recommended starting point
Start with the AI Assurance Evidence Review Kit
Use the general workflow first. Move to the AML Pack when the review requires deeper AML-specific documentation exploration.