AI Reflection Data Handling

Product: AI Dream Reflection

Developer: Dzianis Huletski

Contact: d.h.emservice@gmail.com

Dream text can contain highly personal emotional material. For that reason, AI Dream Reflection is intended to handle reflection-related data with extra caution, clear user-facing language, and minimal transmission wherever possible.

1. Current MVP Status

In the current MVP build, reflection generation is mocked locally. Dream text, optional emotion tags, sleep quality, and draft content are stored only on the user’s device using local app storage. No live AI request is sent in the current offline build.

2. What Data May Be Used for Reflection

If live AI reflection is introduced in a future version, the reflection request may use some or all of the following user-provided inputs:

• dream text entered by the user;
• selected emotion tag;
• sleep quality rating;
• the optional “include context from yesterday” toggle and any related contextual information if later implemented.

3. Sensitive Nature of Dream Content

Dream entries may contain emotional, intimate, distressing, or otherwise sensitive personal information. They may refer to relationships, fears, memories, or private internal experiences. Because of this, dream text should be treated as sensitive user content.

4. Intended Principles for Future AI Processing

If cloud-based AI reflection is added later, the intended handling model is:

• only the minimum necessary text is sent for the requested reflection;
• data is transmitted securely over HTTPS;
• users are informed clearly before external processing is enabled;
• reflection generation is not used for advertising or cross-app profiling;
• data sharing is limited to what is necessary to provide the feature.

5. Transparency Before Sending Data

Before any live AI provider is used in a production release, the app should present clear notice explaining:

• what content is sent;
• why it is sent;
• whether it is processed by a third-party AI provider;
• where users can read the current privacy and data-handling terms.

6. No Hidden Data Use

Dream content should not be silently reused for unrelated purposes. The intended product direction does not include using dream text for advertising profiling, resale, public training disclosure by default, or hidden behavioral targeting.

7. Local-First Preference

The product is conceptually aligned with a local-first or privacy-conscious approach wherever possible. Even if live AI is introduced, local storage and local user control should remain central to the user experience.

8. API Keys and Security

If external AI processing is implemented, API credentials should not be hardcoded directly into a production client app. A safer architecture may include a secure backend or proxy layer to avoid exposing provider credentials in the app binary.

9. Data Retention Expectations

In the current build, dream entries remain on the device until the user deletes them. If future server-based AI processing is added, any transmitted dream content should be retained only as long as technically necessary to process the requested reflection, unless clearly disclosed otherwise.

10. User Control

Users should be able to understand whether reflections are being generated locally or via an external service. Product design should keep this boundary visible rather than hidden.

Users should also retain control over saved entries, including the ability to delete journal entries and drafts stored on-device.

11. Third-Party Provider Considerations

If an external AI model provider is used, that provider may process text according to its own service terms and infrastructure. The app’s legal pages should be updated accordingly before launch of such a feature.

12. Reflection Output Handling

Generated reflections may be stored in the user’s local journal together with the original dream entry. This means both the source text and the AI-produced reflection may remain on-device as part of the journal history unless deleted by the user.

13. No Promise of Anonymity from User Input Alone

Users should avoid entering personal details they would not want processed if live cloud AI reflection becomes available in the future. Even when names are not included, dream text can still be personally sensitive.

14. Future Changes

If live AI reflection is introduced, this page and the Privacy Policy should be updated to reflect the exact provider model, transmission method, retention logic, and any user choices offered.

15. Contact

If you have questions about reflection-related data handling, contact:

Dzianis Huletski
Poland
d.h.emservice@gmail.com