How to Export and Correct an AI Journal Record

Start with the source, not the interpretation

If a generated read seems wrong, reopen the original entry first. Check the date, wording, tags, and any structured fields such as mood, energy, stress, or self-reported sleep hours. A wrong date or accidental rating can change the surrounding context. A summary may also be wrong even when the source is accurate.

Source accuracy matters, but a correction and a generated interpretation are different records. When possible, preserve enough context to compare the inaccurate detail, the requested correction, and any generated read that prompted the review.

Use the documented privacy-request route

Use this form to access, correct, delete, or limit how Daylogue uses your personal information. The form at /privacy/request separates Right to Know, Right to Delete, Right to Correct, Limit Use of Sensitive Personal Information, and other privacy requests so the request can name the control you are seeking. We respond to all verified requests within 45 days (30 days for GDPR).

The public sources for this page do not verify a current export format, file type, menu location, or per-entry correction flow. Older screenshots or instructions may be outdated; check the live product or current support documentation. If a self-service export exists, record its date and the app version with your copy.

Make the correction inspectable

A correction request can identify the smallest item that is actually inaccurate. For example, name the entry date, the value as shown, the accurate value, and why the change matters. If the source is accurate but a generated interpretation does not fit, describe that as an interpretation issue rather than rewriting the source to fit a preferred story.

The public sources do not establish whether a privacy correction changes a journal entry in place, whether generated outputs refresh, or whether historical snapshots remain. List the surfaces you want Daylogue to address in the request, then use the response to distinguish what was corrected from what remains unchanged.

1. 1

   Save the original

   Capture the entry, timestamp, fields, and generated read needed to identify the issue.

2. 2

   Identify the error

   Separate a wrong source value from a poor interpretation of an accurate source.

3. 3

   Correct narrowly

   At /privacy/request, describe the inaccurate text or field, the accurate information, and the reason for the request.

4. 4

   Recheck outputs

   After Daylogue responds, inspect the relevant narratives, candidate patterns, and summaries to see what changed and what stayed fixed.

A correction does not prove a new cause

Correct data can still support several interpretations. A recurring overlap does not show that one thing caused another. If an edited sleep value removes one example from a stress and sleep pattern, the right response may be lower confidence, not a new causal story.

Daylogue's evidence page separates Product Verified, Bench Validated, Human Tested, and Outcome Supported. Product Verified means confirmed through code review and product QA. It does not mean every generated observation is correct for every person. Keep the personal record inspectable and your own interpretation revisable.

Review the export as sensitive data

An access copy or export can concentrate years of personal writing in one file. Store it somewhere appropriate for its sensitivity, check cloud-sharing permissions, and remove temporary copies after the correction is complete. When one entry or screenshot is enough, keep the rest of the archive private.

NIST's Privacy Framework is a voluntary tool intended to help organizations identify and manage privacy risk while protecting individuals' privacy. For a person handling an export, the practical version is simple: know where the copy went, who can access it, why it is needed, and when you will remove it.

What the export and correction diagram shows

The diagram separates three routes: access a copy, correct details, and delete or limit. These are related privacy controls, but they are not interchangeable. An access request asks what information Daylogue holds. A correction request identifies inaccurate information. Deletion or limitation asks Daylogue to remove information or restrict how sensitive personal information is used.

The diagram is a process summary, not a product screenshot. It does not establish a self-service export button, file format, entry editor, or automatic refresh path. The current request form at /privacy/request documents how to submit these privacy requests.

• Verify the current control instead of guessing from the calendar screen.
• Save only the data needed for the correction.
• Keep source edits separate from interpretation changes.
• Recheck downstream outputs after the source changes.