JURNIVAL Back to JURNIVAL

JURNIVAL — AI Transparency Notice

Last updated: November 2026

1. Why JURNIVAL uses AI

JURNIVAL uses generative artificial intelligence systems to create some of the Challenges offered during an Adventure.

The aim is to turn the already planned trip into more personal, contextual and playable activities.

The AI does not plan the trip on its own and does not decide where you should go: it is used to generate game content based on the information about the Adventure.

2. Which Challenges use AI

Not all Challenges are necessarily generated by AI.

The system may use AI-generated Challenges, created dynamically based on the trip context, and system Challenges, used as a fallback or created through predefined logic.

JURNIVAL technically records the origin of AI-generated Challenges through metadata associated with the Challenge.

3. Which model we use

The current AI engine uses Anthropic — Claude Sonnet 5.5, model claude-sonnet-5-5.

Requests are made through the Anthropic API from JURNIVAL's server. The API key is not exposed to the browser.

4. What information is used

To generate a Challenge, the system may provide: trip name; dates; month; number of travelers; preferences; pace; surprise level; destinations and stops; region and country; date of the stop; public local information; slot and characteristics of the Challenge; any free-text notes; titles of the Challenges already present.

Local information may come from public sources such as Wikipedia, Wikidata and OpenStreetMap.

If a destination cannot be found in the search, you can ask JURNIVAL to look for it: the place name you typed and the country you chose are sent to the AI, which suggests the name of the place and the nearest town. Each suggestion is checked against OpenStreetMap (Photon) before it can be added, and you confirm it; the AI never provides coordinates. If nothing is confirmed, the destination can be added by name only, without a position on the map.

5. What information is not sent

In the normal generation process, the following are not sent to the AI model: email address; user ID; display name; avatar; companions' names; Challenge photos; analytics data; numerical Travel DNA scores; geographic coordinates of the destinations.

If you enter personal information in a note used to generate a Challenge, that information may be included in the request sent to the AI service.

6. How a Challenge is generated

The process can be summarized as:

Adventure → trip context → local knowledge → AI → quality check → Challenge.

The generated content undergoes automatic checks relating to, among other things, structure, concrete action, consistency with the destination, local references, duplication and required characteristics.

If the content does not pass the checks, JURNIVAL may regenerate it or use a fallback Challenge.

7. Local knowledge

JURNIVAL uses a Local Knowledge Layer to provide the AI system with public information about the destinations.

The information may include monuments, viewpoints, natural features, neighborhoods, local products, dishes and traditions.

The information used for generation is associated with the Challenge through provenance metadata.

8. AI can make mistakes

The use of artificial intelligence does not guarantee that every Challenge is perfectly accurate.

A Challenge may contain incomplete or out-of-date information, misinterpretations or suggestions that are not available.

JURNIVAL must not be considered a definitive source of information about places, activities, prices, opening hours or safety conditions.

Before following a Challenge, the user must use their own judgment and check the relevant information when necessary.

9. Safety

An AI-generated Challenge does not authorize the user to enter private property, break laws or regulations, access prohibited places, endanger people, damage property or disturb other people.

If a Challenge appears unsafe, illegal or inappropriate, it must not be carried out.

10. Data and Anthropic

JURNIVAL uses Anthropic's commercial API.

According to Anthropic's documentation, for commercial products inputs and outputs are not used to train the models unless specific conditions apply.

The configuration currently used by the JURNIVAL organization provides for 30-day retention, Zero Data Retention (ZDR) not active and user feedback disabled.

Anthropic also states that its commercial API may involve processing in different geographic regions and that retention is subject to the applicable configuration and agreements.

For more information, see the JURNIVAL Privacy Policy.

11. How to recognize an AI-generated Challenge

JURNIVAL is introducing an explicit distinction between AI-generated Challenges and JURNIVAL / system Challenges.

The origin is also recorded technically in the database.

For AI Challenges accepted by the quality-control system, the elements recorded include, among others, AI origin, engine version, AI model used and information about the local facts used.

12. User control

AI is a tool used by JURNIVAL to create game content.

The user keeps control of their Adventure and can, according to the available features, change the itinerary, remove or edit Challenges, regenerate content and choose whether or not to follow a Challenge.

The AI does not make autonomous decisions on the user's behalf outside the features provided by the application.

13. Updates

The model, the generation system, the sources and the quality-control methods may change over time.

This page will be updated when such changes result in a significant change in the way JURNIVAL uses AI or processes users' data.

14. Contact

For questions about the use of artificial intelligence in JURNIVAL:

privacy@jurnival.com

JURNIVAL — Don't just travel. Play it.

Last updated: November 2026