Responsible AI & Ethical Policy

Last updated: June 2026

As AI technologies become integrated into public services and operational workflows, Telali Limited is committed to designing AI systems that are safe, ethical, and transparent.

1. Data Security & Privacy (GDPR Compliance)

We do not use customer or citizen personal data to train public foundation models. We utilize enterprise-grade APIs with zero data retention clauses, or host models locally on-premises within client networks.

2. Factuality & Traceability (Explainable AI)

To prevent model hallucinations, we exclusively deploy Retrieval-Augmented Generation (RAG) models. Systems are configured with strict grounding parameters, requiring the model to cite the exact internal documentation page used to compile the answer.

3. Human-in-the-Loop Governance

We believe AI should assist decision-making, not replace human judgment. No system designed by Telali will perform automated final determinations (such as grant approvals or citizen notices) without explicit human review.

4. Fairness & Bias Prevention

We run rigorous data audits to identify bias in ingestion datasets and construct system prompts that promote fairness and objectivity in generated responses.