1. Our commitments
We treat AI as decision support under human accountability. Every output is attributable to a run record with prompt, logs, timestamps and status so decisions can be audited.
2. Human review and approvals
Runs touching sensitive domains such as health, legal, financial advice, safety, employment and law enforcement are flagged automatically and withheld from clients until a reviewer approves or rejects them. Every decision is recorded in an approval audit log with the reviewer, timestamp, notes and the notification sent.
3. Limitations
Language models can be confidently wrong, can miss context and can reflect biases in their training data. Geospatial and forecasting outputs depend on data quality and vintage. Do not use outputs as the sole basis for consequential decisions.
4. Data used in AI processing
We send only the prompt, agent configuration and relevant requirement context to model providers. We do not use your confidential client content to train public models, and we do not sell it.
5. Prohibited AI uses
Covert surveillance, social scoring, biometric categorisation of individuals, deceptive impersonation, and fully automated adverse decisions about people without human review are out of scope for our platform and services.
6. Transparency and feedback
AI-generated content in the portal is labelled as such. Demonstration data on this site is marked as illustrative. If you believe an output is harmful or wrong, tell us and we will investigate and, where needed, retract it.
Contact: legal@his-ai-agency.com.
This document describes our current practices and is provided for information. It is not legal advice; for a regulated engagement we will agree bespoke contractual terms with you.