How KareBud™ Uses AI
Understanding how AI works in KareBud™, what it can and cannot do, what data it may use, and the choices available to you.
AI as a Mirror, Not a Megaphone
KareBud™ uses AI to organize and reflect the information you choose to enter. This can help you see possible patterns and prepare thoughtful questions without making medical decisions for you.
AI Helps You:
- ✓Summarize what you recorded
- ✓Reflect on patterns across time
- ✓Prepare questions and insights
- ✓Organize your health story
- ✓Label reported facts separately from generated review cues
AI Does NOT:
- ✕Diagnose medical conditions
- ✕Predict disease outcomes
- ✕Recommend treatments or medications
- ✕Replace healthcare professionals
- ✕Verify that a reported fact is clinically correct
Privacy Safeguards
Your health data is yours. Here's how we protect it when AI is involved:
Model Training Controls
KareBud does not intentionally use identifiable health content to train general-purpose models. Contractual terms and configurations vary by provider and feature; see the Privacy Policy for details.
T.R.U.S.T. Proxy Routing
Our trust controls attempt to reduce direct identifiers and route requests according to the selected feature. Automated redaction is not perfect and may miss sensitive details.
Strict Guardrails
AI is strictly used to provide contextual language (such as summarization and sensemaking) grounded in your reported logs. KareBud uses a dual-layer approach combining deterministic rules and probabilistic models with confidence indicators. While these safety guardrails are robust, automated outputs should always be reviewed to ensure accuracy.
User-Controlled Processing
You decide which optional AI features to enable. Pausing AI may limit summaries and reflections, but it does not delete your underlying reported data.
Explainability & Review
What You Should See
When KareBud summarizes or highlights a possible pattern, it should preserve the source meaning and, where available, point back to the reported entry, uploaded record, consent, or correction history that informed the output.
What AI Does Not Prove
A pattern, cue, score, or summary does not prove cause, diagnosis, severity, prognosis, medication safety, or treatment need. Use it as a conversation starter and verify against source records and qualified clinical judgment.
Correction Traceability
If you confirm a correction, KareBud may update the current source record used by active views while keeping before-and-after values in an audit trail. Historical entries are not rewritten to pretend the correction existed earlier.
Clinician View Boundary
Normal user views should not show diagnostic hypotheses. Authorized clinician workflows may show unverified review cues so the clinician can independently evaluate, decide, and document.
Your Controls
Opt Out Anytime
You can disable AI-powered features at any time from your settings. Your data remains accessible in its raw form, and you can reactivate AI features whenever you choose.
Review Consent Records
Settings should show the latest consent records for active AI and data-sharing features. Older consent versions may remain in audit history but should not be presented as the current active consent.
Export & Delete
You can request an export of available account data. Account deletion begins a deletion or de-identification process, subject to limited legal, security, audit, and backup retention.
Feature Consent Notifications
If a feature requires consent that has not been captured, KareBud may block or limit the feature and send an in-app notice asking you to review the required consent before continuing.
Important Notice
KareBud™ does not provide medical advice. Always consult qualified healthcare professionals for medical decisions, diagnosis, or treatment.
AI-generated insights are unverified summaries or review cues based on recorded information. They are not clinical assessments and should not be used as a substitute for professional medical care.
Normal user views receive reflections, summaries, possible patterns, and questions to discuss, not diagnostic hypotheses. Separate clinician-review workflows may show authorized clinicians unverified review cues or hypotheses. The clinician must independently evaluate the source information, decide what it means, document through an approved workflow, and act.
Diagnoses, conditions, procedures, symptoms, and other health facts already reported by you or documented by a clinician may be stored and summarized with a clear source label. KareBud should preserve those facts without hiding them or changing their meaning or intensity. This is different from KareBud generating a new diagnosis, which it does not do.
KareBud does not monitor or reliably detect emergencies. In the United States, call 911 for immediate danger or 988 for a mental health crisis.
By enabling AI features, you consent to the processing described in this notice, the Privacy Policy, and any feature-specific consent shown in the product. You can withdraw consent, but withdrawal may disable the related feature and does not erase audit records that must be retained for traceability, security, or legal reasons.
Questions?
If you have questions about how KareBud™ uses AI or your data privacy rights, please contact us.