Responsible Artificial Intelligence Policy

Kairos (Kairos Group, S.A. (in formation)) · Version 1.0 · Effective: October 6, 2026 · Pending legal review · Courtesy translation: the Spanish version prevails.

Aligned with Panama’s National AI Strategy and international good practice (OECD principles, NIST AI RMF, ISO/IEC 42001).

Our 8 commitments

1. Explicit rules. Every automation has a decision matrix written and approved by the client. The AI does not invent rules.
2. Human in the loop. Whatever does not meet clear criteria goes to human review. No decision affecting people’s rights is left solely in the hands of the AI.
3. Verification afterwards. Every bot action is checked after it runs and is audited.
4. Privacy by design. Sensitive data is processed on the client’s infrastructure, minimized and, where possible, anonymized.
5. We do not train on your data. Neither our own models nor third-party ones.
6. Transparency. We document which models we use, for what and with what limits. Conversational agents identify themselves as automated assistants.
7. Fairness and quality. We test with fictitious cases (“seeded tests”) and measure errors before and after going to production.
8. Accountability. Every project has an owner at Kairos and another at the client, with an incident response plan.

Uses we do not accept

• Surveillance or profiling of people without a legal basis.
• Impersonation or deceptive content.
• Evading security controls, captchas or third-party terms.
• Automated decisions on credit, employment or health without human review.

Risk classification of each project