Responsible AI Policy
Last updated: July 19, 2026 · Governs every model, agent and platform in the OMS AI Core ecosystem — OMS AI Nexus, Developer AI OS, Mother AI, AI Factory and connected products.
1. Purpose
OMS IT™ builds AI systems that augment human capability. This policy sets the principles and controls we apply across design, training, deployment and operation of our AI platforms.
2. Core Principles
- Human oversight — humans remain accountable for consequential decisions; AI recommends, humans decide.
- Fairness — we test for and mitigate demographic, linguistic and cultural bias.
- Transparency — users are told when they interact with AI (e.g. OMS Asking assistant) and what data it uses.
- Privacy by design — minimum-necessary data, encryption in transit and at rest, no sale of personal data.
- Safety — guardrails against harmful, illegal or deceptive outputs; red-team review before release.
- Accountability — every AI system has a named owner inside OMS IT™.
3. Data Use for AI
Client and user data are used only for the purposes agreed in the applicable contract. We do not train foundation models on confidential client data without written consent. Fine-tuning on client data, when contracted, remains isolated to that client's tenant.
4. Model Sourcing
OMS AI platforms combine proprietary systems with vetted third-party foundation models. We evaluate each provider for security, data-handling and jurisdictional posture before integration.
5. Prohibited Use
OMS AI systems must not be used to generate disinformation, non-consensual imagery, malware, discriminatory profiling, unlawful surveillance, or automated decisions affecting legal rights without human review.
6. AI Assistant Disclosure
The OMS Asking assistant on this website is an AI system. It may make mistakes; verify important information with a human at info@omsit.com.
7. Incident Response
Suspected AI-related harm, bias or data incidents can be reported to info@omsit.com. Reports are triaged within one business day.
8. Regulatory Alignment
This policy is designed to align with the Bangladesh ICT Act, the Digital Security Act 2018, and emerging national AI guidance, and to remain compatible with EU AI Act and OECD AI Principles for cross-border engagements.
9. Review Cadence
This policy is reviewed at least annually and whenever a material change is made to an OMS AI platform.