AI Audit, Assessment & Testing
WAIG’s AI Testing Lab delivers not-for-profit audit and assessment services for organisations adopting AI. Our practitioners run structured test catalogs, red-team scenarios, and continuous monitoring to validate safety and robustness — with findings documented for governance boards and regulators.
The AI Testing Lab is WAIG Foundation’s independent assurance vehicle for model and system risk. Engagements combine structured test catalogs (adversarial, fairness, robustness, and misuse scenarios) with practitioner-led red-teaming and evidence packs suitable for boards, auditors, and regulators.
Unlike vendor product certification, Lab work is advisory and public-interest oriented: findings map to NIST AI RMF, OWASP LLM Top 10, and ISO/IEC 42001 control expectations without locking clients into commercial software.
Key capabilities
- ·Adversarial & red-team catalogs (prompt injection, jailbreaks, data leakage)
- ·Deepfake / media authenticity and misuse scenario testing
- ·Fairness, bias, and robustness validation with documented evidence
- ·Model drift and continuous monitoring recommendations
- ·Board- and auditor-ready findings reports
Who it serves
- CISOs and security teams
- AI product and MLOps owners
- Risk, compliance, and internal audit
- Public-sector and regulated industries
Engagement deliverables
- ✓Scoped test plan aligned to use-case risk
- ✓Executed adversarial / robustness / fairness assessments
- ✓Evidence pack with reproducible findings
- ✓Remediation roadmap and retest guidance
Outcomes
- →Clear residual risk picture before production or certification
- →Audit-ready documentation for ISO 42001 and sector regulators
- →Stronger human oversight and incident response readiness
Standards alignment
- NIST AI RMF
- OWASP LLM Top 10
- ISO/IEC 42001