Rules for the use of recommendation technologies

1. General

1. These Rules are published under Article 10.2-1 of Federal Law No. 149-FZ of 27 July 2006 "On information, information technologies and information protection" and describe which recommendation technologies the Frolick service (the "Service") uses, what information they rely on and how recommendations are formed.
2. Recommendation technologies are information technologies that present information based on collecting, organising and analysing data about user preferences. The operator of the Service is named on the About the operator page.

2. Where recommendations are used

• Community catalogue (public games, content sets, lessons and boards): ordering the list by popularity, recency and match with the selected filters, and the "Similar materials" block on a material's page.
• Workspace: suggestions while creating a game or lesson — suitable templates, content sets and game types for the chosen subject, age group and language.
• AI assistant: content options (questions, words, tasks) proposed from your request and the materials in your workspace. Such content is labelled as AI-generated in line with the AI usage rules.

3. What information is used

1. Attributes of the materials themselves: subject, topic, language, age group, game type, tags, publication date, number of views, copies and ratings, average rating.
2. Your current request: the search string, selected filters and sort order, the material you have open.
3. Data from your workspace — only for suggestions and the AI assistant inside that workspace: the materials you created, the chosen subject, interface language and settings.
4. Anonymised statistics on use of the Service's features.

4. How recommendations are formed

1. Catalogue ordering uses formulas that weigh the match with filters and the query, the material's recency and its popularity (views, copies and ratings over the recent period). Materials are not promoted manually, except editorial picks labelled as such.
2. "Similar materials" are selected by matching subject, tags, game type and age group with the open material.
3. Workspace suggestions and AI assistant answers are produced by a language model from your request and your workspace materials; they are advisory and do not replace the teacher's judgement.
4. Personal interest profiles that combine a user's behaviour across sessions and accounts are not built. Catalogue recommendations are identical for every user with the same query and filters.

5. Controlling recommendations

1. Catalogue sorting can be changed at any time (by date, alphabetically, by popularity); the "Similar materials" block and suggestions are optional and do not affect access to the Service's features.
2. The AI assistant can be switched off in workspace settings; without it, no suggestions are generated from your materials.
3. The data listed in section 3 is processed in accordance with the Personal Data Processing Policy and the Cookie Policy.

6. Changes and contacts

1. When new recommendation technologies appear in the Service, these Rules are updated before they are put into use.
2. Questions about the Rules: support@frolick.ru or the Contact page.