⚡ Direct Answer Summary (AI Overview / Key Takeaway)
Terms governing client engagement with ThreatKrusher AI capabilities, prompt isolation boundaries, and model usage terms.
eTrepid ThreatKrusher AI Service-Specific Terms (Schedule)
Governing Standard: NIST AI RMF / Commercial Agreement Hierarchy
1. Structure and Applicability
These AI Service-Specific Terms ("AI Terms") form a binding Schedule to the governing eTrepid Master Service Agreement (MSA) and Statement of Work (SOW) whenever the customer subscribes to or utilizes eTrepid AI-assisted capabilities, including AI-as-a-System, Comply AI compliance acceleration, or automated SOC triage features.
flowchart LR
In["Customer Inputs & Prompts"] -->|Customer Property| Eng["ThreatKrusher AI System"]
Eng -->|Customer Property| Out["Generated Compliance Deliverables & Outputs"]
Customer Inputs: The customer retains sole ownership of all text, system specifications, architecture notes, and prompts provided to the AI system ("Inputs").
Customer Outputs: Between the parties, the customer owns all deliverables, customized policies, compliance mappings, and analysis generated by the AI system in response to customer Inputs ("Outputs").
No Public Retraining: eTrepid represents and warrants that Inputs and Outputs will not be used to train, retrain, or fine-tune public machine learning foundation models.
3. Human Oversight & Operational Limitations
Informational & Decision Support: Outputs generated by AI tools are designed for expert decision support and operational acceleration. Outputs do not constitute binding C3PAO audit approvals or formal legal advice.
Client Validation Duty: The customer agrees that high-consequence business, contractual, and technical decisions based on AI outputs must be reviewed and validated by qualified personnel prior to formal submission or execution.
4. Sovereign Enclave Operation
Where services involve defense compliance data or Controlled Unclassified Information (CUI), AI processing pipelines execute within FedRAMP High / DoD IL4-authorized enclaves (e.g., Azure Government OpenAI Service). Access is strictly restricted to authorized, US-based infrastructure.
