Fictional-scenario demo and information sources
This independently developed public demo is for adults aged 18 or over testing fictional adult scenarios. It is not a company, hospital or real healthcare service. Do not enter real health information or identifiers, including in files. Submitted showcase cases may be available to authorized demo doctor accounts in the same configured demo hospital; those accounts do not provide healthcare. User text is not assumed anonymous or synthetic merely because this is a demo.
A local privacy gateway reduces known account identifiers and recognizable identifiers in text before AI requests, including conversation replay. It is not infallible anonymization. OpenAI store:false does not by itself disable all provider retention or establish HIPAA compliance.
Approved general health explanations may use permitted MedlinePlus health-topic summaries. Source details are available in the results section. General descriptions are separate from the AI’s personalized, uncertain assessment.
01
Who operates this public demo
Alhan Akdemir is the owner and developer of SafeSymptom and the individual operator/data controller for the processing described in this notice. SafeSymptom is an independent personal project, not a healthcare organization. The operator determines processing purposes, permissions and application access.
02
Data we receive and where it comes from
- Google account identifier, name, email, and sign-in events from you and Google OAuth.
- Fictional scenario inputs you provide: age range, biological sex/pregnancy context, allergies, conditions, medications, symptoms and chat responses. If real health information is entered despite the warning, it can still be processed and requires appropriate protection; the warning does not remove that responsibility.
- AI questions, risk classifications, possible-condition explanations, and clinician summaries generated from your assessment.
- Demo doctor messages, consultation-flow records, attachments, simulated prescriptions, lab orders, follow-ups and appointments when those demo features are used. These are not real healthcare records or services.
- Security and reliability events such as session, route, rate-limit, and redacted error information from our systems.
03
Why we process it
We process account and interaction information to authenticate users, generate fictional-scenario AI responses, apply safety rules, display demo results and consultation flows, send selected or necessary account/workflow messages, secure and debug the demo, meet applicable legal obligations and answer privacy requests. Separate demo-evaluation permission covers the limited developer review described below.
We do not sell health data, use it for targeted advertising, or use health-data geofencing. We do not intentionally add your name or email to AI prompts. Identifiers you type into free text can still reach the active AI provider.
Separate permission for developer review
Participation requires an unchecked, separate demo-evaluation choice allowing the owner and developer of SafeSymptom to review new test messages, AI responses and related interaction records for product quality, safety and usability. The information-notice acknowledgement and provider-processing consent remain separate, and email reminders remain optional. This is a participation condition for this test demo, not a legally validated consent arrangement or a waiver of legal rights.
Review is limited to tests started after the new permission was recorded and while permission remains current. A new grant does not cover older conversations, including conversations from before withdrawal and re-consent. Withdrawal ends permission for future evaluation use, including review of previously eligible records. No automated evaluation pipeline, model training or new review dashboard is enabled by this update. This permission does not authorize public sharing or transfer to third-party evaluation programs such as IBM.
04
Retention and deletion
This status reflects the application mode and operational switch, not a guarantee that every scheduled run succeeds. Paused or failed work can delay deletion.
Newly enrolled demo conversations become eligible for scheduled cleanup 30 × 24 hours after the linked group’s last meaningful interaction. Messages, completed AI responses, attachments and saved consultation actions renew the period; viewing does not. Existing records, their continuations and fixed synthetic datasets are not reclassified.
Account, security, consent evidence and rights-request records have separate purposes. No verified automatic expiry is promised here.
Linked transfers are cleaned together, including conversation content, encrypted operation inputs/responses, attachments and related demo records. Active model or upload work, calls or room cleanup, outstanding appointments/follow-up rights, and links to records outside the group defer cleanup. An open chat alone does not prevent expiry. Once deletion starts, that conversation cannot receive new content; start a new demo instead. Independent patient history is not added to the deletion scope merely because a doctor viewed it.
You can submit and track deletion requests in the Privacy Center. Withdrawal alone does not delete records. Application conversations, storage objects, backups and provider logs do not share a single deletion schedule. Provider retention depends on the applicable service and configuration; exact backup and provider-log deletion times have not been verified for this demo. Necessary evidence may be retained where an applicable legal obligation permits or requires it; the response to your request should explain any exception.