Assessment Disclaimer

A MaacVerify assessment report is an independent, evidence-based assessment of how an AI system performed against a defined scenario corpus within a specified deployment domain at a point in time.

It is not a guarantee of safety, a regulatory approval, a legal opinion, or a assessment of the model in general. It is a fitness-for-purpose assessment for a bounded deployment context.

Every assessment is bounded by domain specificity, corpus specificity, point-in-time validity, and exact version specificity. Claims about a assessed system must not exceed the assessed domain, corpus, date, or system version.

Deployment suitability guidance indicates fitness for purpose under assessed conditions. It does not guarantee live deployment outcomes and does not cover human factors, workflow integration, security, prompt-injection resilience, or adversarial misuse unless expressly stated.

A MaacVerify assessment may not be represented as a safety assessment, deployment approval, regulatory clearance, legal opinion, or blanket endorsement of a system across all use cases.

References to assessment should remain qualified by domain, scope, assessment date, and system version. Unqualified claims that a system is simply 'MAAC assessed' risk overstating the report's meaning.

Reliance on a assessment should be revisited when the model version changes, prompts or system instructions change, the deployment domain expands, monitored drift exceeds tolerance, the governing regulatory framework changes materially, or an adverse incident occurs.

MaacVerify recommends annual reassessment for systems deployed in regulated environments even if no material trigger has occurred.

MaacVerify assessments are evidentiary artifacts intended to support procurement, governance, and compliance programs. They do not by themselves establish legal compliance with the EU AI Act, HHS guidance, FINRA, OCC, FDA frameworks, or any other law or regulation.

Organizations remain responsible for their own compliance determinations, deployment decisions, and legal analysis. Nothing on this site or in a assessment report constitutes legal, clinical, financial, or regulatory advice.

Current MAAC v4.7 assessments are subject to known methodology limits, including equal weighting across the nine dimensions, corpus-bounded generalization, model-administered adjudication, and evolving domain-extension validation.

Assessment reflects performance at the time of assessment. Future framework revisions, deployment changes, or environmental drift may change the meaning of a prior result.

MaacVerify is an independent assessment authority. It does not build, sell, train, or distribute AI models and has no financial interest in the commercial success of the systems it assesses.

Assessments are commissioned by the organization accountable for the deployment, not by the model vendor. This structure is intended to remove the conflict of interest inherent in vendor self-evaluation.

If there is any conflict between documents, the executed Statement of Work governs the specific engagement first, followed by the Assessment Disclaimer, then the Terms of Service, and then the Privacy Policy.

Questions regarding the scope, limitations, or appropriate use of a MaacVerify assessment report may be sent to info@maacverify.ai. Attorney review is recommended before production reliance on new disclaimer revisions.