Gentoura AI Ethics & Transparency Charter

At Gentoura, our product vision is built on Amplified Directness—collapsing the traditional, cluttered travel booking experience into intuitive, single actions. A central component of this experience is our AI Recommendation Assistant. We believe that integrating artificial intelligence into travel discovery requires an unwavering commitment to ethics, transparency, and user trust.

This Charter details our ethical framework and the operating constraints under which our AI systems run.

1. Core Principles

Our AI strategy is guided by three core design and business principles:

  1. Transparency of Capability: We are explicit about what our AI can and cannot do. We do not represent our conversational discovery assistant as a human agent.
  2. User Autonomy: The AI acts as an assistant and filter, not an automated decision-maker. The user always retains full control over selection, booking decisions, and final details.
  3. Data Minimization & Ephemerality: We only collect the minimal text input required to recommend travel itineraries, and we do not build persistent, invasive behavioral profiles.

2. Scope & Functionality of the AI Assistant

The Gentoura AI Recommendation Assistant is designed strictly to facilitate the discovery of existing tour options.

  • Fixed Package Matching: The AI is programmed to scan our curated database of pre-defined, fixed tour itineraries (Static Packages). It does not dynamically assemble custom itineraries, invent flight schedules, or dynamically change package prices.
  • Parameter Extraction: The AI processes natural language inputs to extract key criteria (such as destination countries/cities, available budget, preferred trip duration, and travel themes like heritage, adventure, or cruises) to match you with the closest tour package in our system.
  • No Autonomous Purchasing: The AI cannot finalize bookings, execute financial transactions, or charge cards. It serves exclusively to recommend tours and direct the user to details pages.

3. Transparency & Prevention of Hallucinations

We recognize the risk of generative AI "hallucinations" (generating false or misleading information). To protect our travelers, we have put the following safeguards in place:

  • Pre-Defined Source of Truth: The AI recommendation engine is grounded entirely in the active database of TourResource entries. It is not permitted to suggest destinations or accommodations that Gentoura does not offer.
  • Dynamic Data Safeguards: Real-time pricing, availability, and specific accommodation tiers are populated directly from the database to the Tour Details page, rather than generated by the AI text handler.
  • Fail-Safe & Graceful Fallbacks: If the AI is unable to find a clear match within our catalog, or if system token limits are exceeded, the assistant is constrained to display popular tour listings and recommend using standard category filters, rather than inventing placeholder packages.

4. Privacy & Ephemeral Interactions

To ensure absolute user privacy in our MVP phase:

  • Guest-Only Model: Because Gentoura operates on a guest-only model (no user accounts or persistent profiles), AI chat sessions are transient. They are tied to temporary session cookies and are cleared once the browsing session is closed.
  • Input Constraints: The AI Assistant accepts a maximum text length of 1,000 characters. We advise users not to share sensitive personal details, passports, or payment details in the open-text chat bubble.
  • No Long-Term Logging: Conversations are processed in real-time. We do not use user chat logs to train general external LLM models.

5. Human-in-the-Loop Oversight

We believe technology should amplify, not replace, high-quality human hospitality.

  • Offline Verification: All bookings initiated on the site are received as offline reservation requests.
  • Human Concierge Review: A professional Gentoura operations specialist manually reviews every request, contacts the traveler to confirm details, and handles payment processing securely offline. This guarantees that no automated AI recommendation can result in accidental bookings or unauthorized charges.

6. Feedback & Recourse

We are continuously refining our AI system's performance and accuracy. If you receive an inaccurate recommendation or encounter unexpected behavior:

  • Use the Contact Us page to submit a direct message.
  • Our operations team will investigate and refine the system parameters accordingly to prevent future errors.