KokuaOS — Responsible AI Policy

KokuaOS's approach to responsible AI and the human-oversight, transparency, and high-risk-use expectations that apply to everyone who uses the AI Services.

Capitalized terms have the meanings in the KokuaOS Definitions Schedule (kokuaos.com/legal/definitions).

1. Purpose and approach

The Services use artificial intelligence to power AI Employees across voice and messaging Communication Channels. This Responsible AI Policy describes how KokuaOS approaches AI responsibly and the responsibilities that Customers, Service Providers, Resellers, and their Authorized Users share when they configure and operate the AI Services. It is a Policy incorporated by reference into the Agreement and works together with the Acceptable Use Policy (kokuaos.com/legal/aup) and the Compliance & Shared Responsibility Policy (kokuaos.com/legal/shared-responsibility).

KokuaOS designs the Platform with configurable safeguards — including the Compliance Shield and guardrail, disclosure, and access-control features — intended to support responsible, human-supervised use of AI. These features support responsible use; they do not replace human judgment, and they do not guarantee any particular outcome. Responsible outcomes depend on how each Customer configures, supervises, and uses the AI Services.

2. AI Output and its limitations

AI is probabilistic and imperfect. The following clause from the Agreement governs AI Output and applies in full under this Policy:

AI Output; Human Oversight. The Services use artificial intelligence. AI Output may be inaccurate, incomplete, outdated, biased, or otherwise unsuitable, and may "hallucinate" or produce content that appears authoritative but is wrong. AI Output is not professional, legal, medical, financial, or other regulated advice. Customer is solely responsible for reviewing AI Output and for any decision or communication it enables, and must maintain qualified human review before relying on AI Output for any legal, medical, financial, employment, safety, eligibility, or other high-risk or legally significant decision. KokuaOS disclaims all liability for reliance on AI Output to the fullest extent permitted.

The AI Services provide assistance only and do not replace human judgment. In addition to the limitations above, AI Output may "hallucinate" or fabricate information — inventing facts, citations, names, quotes, or events that are false or never occurred — omit material facts, be internally inconsistent or return different results for the same or similar inputs, be outdated, and misunderstand or misinterpret prompts, instructions, or context. The Customer must review important, consequential, or legally significant AI Output before relying on it. KokuaOS does not warrant that AI Output will be accurate, complete, current, or error-free, and does not guarantee that the AI Services will be uninterrupted or that safeguards will catch every problematic input or output.

3. Human oversight

The AI Services are designed to assist and augment people, not to operate unchecked. The Customer must maintain meaningful human oversight proportionate to the risk of its use, including the ability to monitor, review, correct, escalate to a human, and disable or override an AI Employee. The Customer is responsible for defining escalation paths (including transfer to a human where appropriate), reviewing AI Output before it is relied upon for consequential purposes, and supervising the outcomes of its AI Employees.

4. Transparency and disclosure

Where disclosure that a person is interacting with artificial intelligence is required by law, or is appropriate to avoid deception, the Customer must configure and deliver that disclosure — for example, telling callers and messaged parties that they are interacting with an AI Employee, and identifying the business on whose behalf it operates. KokuaOS provides configurable disclosure features to support this; the Customer is responsible for determining what disclosures apply to its use and for enabling them. The Customer must not use the AI Services to deceptively impersonate a person, or to misrepresent that AI Output is human-generated, in violation of law or the Acceptable Use Policy.

5. High-risk uses

Certain uses carry heightened risk to individuals' rights, safety, finances, or legal position. For any high-risk or legally significant use — including decisions or communications affecting legal rights, medical or health matters, financial or credit outcomes, employment or hiring, housing, insurance or eligibility, or public or personal safety — the Customer must not rely on the AI Services to make or deliver autonomous decisions without qualified human review. A human must remain responsible for the decision. The Customer must also comply with any law that restricts automated decision-making or requires human review, notice, explanation, or an appeal or opt-out right. KokuaOS may decline to support, or may restrict, uses it reasonably determines present unacceptable risk, are prohibited by the Acceptable Use Policy, or are not approved for the applicable data or workflow.

6. Bias and fairness

AI systems can reflect biases present in data, prompts, or configuration and can produce inconsistent results across individuals or groups. KokuaOS works to design and select AI capabilities with fairness in mind, but does not warrant that AI Output will be free of bias. The Customer is responsible for evaluating its own configurations, prompts, and Knowledge Base content for fairness and legal compliance in its context, for testing outcomes for its population and use case, and for meeting any anti-discrimination and equal-treatment obligations that apply to its use.

7. Access controls

The Customer is responsible for administering access to the AI Services, including provisioning and deprovisioning Authorized Users, assigning roles and permissions on a least-privilege basis, protecting credentials, enabling available authentication controls, and restricting who can configure, deploy, or modify AI Employees and their prompts, workflows, and integrations. Additional detail is in the Security Policy (kokuaos.com/legal/security).

8. Testing and monitoring

KokuaOS tests and monitors the Platform and its AI safeguards on an ongoing basis and may adjust models, safeguards, and configurations to improve safety, quality, and reliability, consistent with the Agreement. The Customer is responsible for testing its own AI Employees before Production use and for monitoring them in Production. Testing and Sandbox activity may generate billable AI Consumption and Variable Charges, as described in the AI Services & AI Consumption Policy (kokuaos.com/legal/ai-consumption). Measured usage records govern; dashboards and alerts are estimates and do not cap or guarantee cutoff.

9. Customer responsibilities

Each Customer (and each Service Provider and Reseller for its own and its downstream use) is responsible for:

• configuring, supervising, and using the AI Services lawfully and in accordance with this Policy, the Acceptable Use Policy, and the Compliance & Shared Responsibility Policy;
• maintaining qualified human oversight and review appropriate to the risk of its use;
• providing legally required AI and recording disclosures and obtaining required consents;
• ensuring its prompts, Knowledge Base content, and workflows are accurate, lawful, and free of prohibited data;
• not using the AI Services for autonomous high-risk decisions without human review; and
• training its Authorized Users on responsible use and overseeing its downstream Customers' use.

10. Incident reporting

If you become aware of AI behavior that causes or is likely to cause material harm — such as unsafe, discriminatory, deceptive, or seriously inaccurate AI Output, a safeguard failure, or a suspected misuse of the AI Services — promptly report it to KokuaOS at trust@kokuaos.com (or the contact in the Trust Center, /trust), and take reasonable steps to mitigate, including pausing or disabling the affected AI Employee where appropriate. Security incidents are handled under the Incident Response Policy (kokuaos.com/legal/incident-response) and suspected vulnerabilities under the Vulnerability Disclosure Policy (kokuaos.com/legal/vulnerability-disclosure). The Customer remains responsible for any notifications to regulators or affected individuals that its use requires.

11. Relationship to the Agreement

This Policy is incorporated into and forms part of the Agreement, and its defined terms have the meanings given in the Definitions Schedule. Nothing in this Policy expands KokuaOS's obligations beyond the Agreement or limits the disclaimers, limitations of liability, or indemnities in the Agreement, which continue to apply and which protect the KokuaOS Parties as intended third-party beneficiaries. This Policy states responsibilities and expectations for responsible use; it does not create any warranty, guarantee of error-free output, or third-party right beyond the KokuaOS Parties as stated in the Agreement.