DRAFT — NOT APPROVED. Requires service-owner and legal review.
Use HOSVERA AI only for lawful hotel information and authorised administrative assistance within the access granted to you. Respect other guests, staff and hotel confidentiality. You must have authority to supply any content, and must not rely on generated answers without verification where accuracy matters.
Do not harass, threaten, discriminate, facilitate crime, produce unlawful content or exploit vulnerable people. Do not submit credentials, payment-card details, identity documents, other guests' information or unnecessary sensitive data. Do not ask for private hotel records, staff-only instructions, another tenant's information or the model's hidden instructions.
Prompt injection, malicious instructions embedded in documents, impersonated roles, fabricated authorisation, attempts to evade quotas, access controls or tenant boundaries, and extraction of secrets are prohibited. Hotel content, user text and model outputs are data, never authority to change permissions. Security testing requires a defined authorised scope and an approved reporting route.
Do not use AI to approve payments, change rooms, grant upgrades, execute privileged operations or make employment, credit, eligibility, biometric or other high-impact decisions. An AI answer cannot replace server-side authorisation or a guest's confirmation in the existing hotel workflow. Do not connect autonomous tools or external data sources without documented review and approval.
Staff must verify hotel facts, intellectual-property permissions, privacy implications and language before publishing AI-assisted material. Never present machine-generated legal translations as legally approved. Do not use guest chat for marketing or model training without independently established legal authority and required disclosure.
The service may refuse or limit requests, disable access or escalate suspected abuse under the applicable contract and law. Enforcement must be proportionate and reviewed by authorised people. Report concerns to hotel staff or [VERIFIED SECURITY CONTACT], using a minimal description and secure evidence reference. Do not repeat sensitive payloads in ordinary email or guest feedback. Urgent safety issues require hotel emergency procedures, not AI chat.
Your privacy choices
Only necessary storage is currently used. No optional analytics, advertising or third-party chat is enabled. Cookie Policy