OmegaOS
AI governance

AI Use and Accountability Policy

Public rules for AI-assisted work, provider routing, human oversight, evidence, and prohibited reliance.

Status
Published policy
Owner
Omega AI Governance and Security
Effective
July 21, 2026
Last updated
July 21, 2026
Internal review posture

Published following operator-authorized Omega privacy, security, claims, and operational review. External counsel review is not recorded.

External counsel posture
External counsel review recommended

AI Use and Accountability Policy is operator-authorized for public publication. External counsel approval is not recorded and remains recommended before material commercial or regulatory reliance.

Purpose

Omega uses AI to assist with research, drafting, analysis, orchestration, retrieval, content, and governed execution. This policy explains the controls and responsibilities that apply when AI contributes to work.

Human authority

AI does not replace the legal or operational authority of the customer, user, or designated approver. Material actions should have an identified owner, permitted scope, review posture, and stop condition appropriate to risk.

Provider and model routing

Omega may route a task to local or third-party models based on capability, configuration, cost, availability, privacy, and risk. The selected route can vary by workflow. Customers must not assume a specific provider unless the agreement or product setting identifies one.

Inputs and Customer Content

Users must have the right to submit inputs and must apply appropriate data minimization. Do not submit secrets, regulated data, or personal information to a workflow unless the service, agreement, permissions, and provider configuration support that use.

Training posture

Omega does not intentionally train an Omega-owned general-purpose model on Customer Content without separate agreement. Third-party training and retention terms depend on the selected provider and account. Omega uses contractual and configuration controls where available but does not rewrite a provider's terms through this policy.

Output limitations

AI outputs can contain factual, legal, security, bias, attribution, or intellectual-property errors. Users must verify material claims and use qualified review for consequential decisions. An explanation, score, citation, or system trace does not by itself prove correctness.

Evidence and traceability

Governed workflows may record prompts, sources, route decisions, tool calls, approvals, outputs, costs, and evidence to support security, replay, accountability, billing, and learning. Recording depends on the feature, retention setting, and lawful purpose.

Prohibited AI use

AI may not be used through Omega for unlawful surveillance, exploitation, impersonation, deceptive manipulation, unauthorized access, rights violations, weapons harm, or high-impact decisions without lawful authority and appropriate human safeguards.

Synthetic and public content

Public-facing AI-assisted content must follow Omega claims, brand, privacy, intellectual-property, and review requirements. Synthetic media must not mislead people about material facts or identity.

Incidents and questions

Report unsafe output, suspected data exposure, or unauthorized AI activity to security@omeganeural.com. Policy questions can be sent to legal@omeganeural.com.

Related policies

Read this page with the related policy, trust, and service information that applies to your use of Omega.

Contact Omega