Data Collection
- Authentication: Email address (Firebase Auth UID) to create and secure accounts, support password recovery, and prevent duplicates.
- Profile Information: User-chosen first name, anonymized last-name noun, profile photo (optional), and “About Me” text to personalize the experience and support sponsor matching.
- Chat Transcripts: Peer-to-peer and AI chatbot dialogs—fully de-identified before analysis—to refine support algorithms and measure engagement.
- Usage & Activity: Feature taps, session timestamps, task completions, daily active usage, and app-open frequency to identify high-value features and monitor engagement.
Data Usage
Collected data is used to support both application functionality and ongoing research into digital recovery support tools:
1. Research and Analytics: De-identified and aggregated data may be analyzed using natural language processing and machine learning techniques to better understand factors associated with recovery engagement, sustained sobriety, and relapse risk.
2. Service Improvement: De-identified user interactions and feedback may be used to evaluate and improve the quality, relevance, and effectiveness of Sponserity features, including the Dawn AI recovery companion.
Personally identifiable information is not used for research analysis without appropriate consent, and research datasets are de-identified prior to analysis whenever feasible.
Data Storage & Security
- Primary Storage: Firebase Realtime Database / Firestore & Cloud Storage.
- LLM Inference: OpenAI API, configured not to train on submitted content—no transcripts are stored on their servers.
- Encryption: Google Cloud default at-rest encryption; TLS/HTTPS in transit.
- Access Controls: Production-DB credentials held only by authorized team members; principle of least privilege enforced.
Confidentiality & De-identification
- UIDs: Random, non-sequential internal identifiers replace personal emails/names in research exports.
- Automated Scrubbing: OpenAI-powered pipeline removes phone numbers, addresses, last names, and other PII from transcripts.
- Timestamps: Original message times retained to preserve behavioral patterns (IRB-approved).
- Quality Checks: Manual spot-checks ensure no residual PII remains before analysis.
Participant Rights
- Voluntary Participation: Use of the app is entirely optional and may cease at any time without penalty.
- Withdrawal: Account deletion via Settings → Delete Account or by emailing support@sponserity.com halts new data collection; existing data remains in fully de-identified form and UID mappings are destroyed.
- Data Access: Requests for personal data exports fulfilled within 30 days via support@sponserity.com.
- Data Correction: Profile updates may be made in-app under Settings → Profile.
- Complaints: Reports of rights violations directed to support@sponserity.com or the University IRB Office at muresearchirb@missouri.edu, (573) 882-3181.
Contact
Privacy Contact: Sponserity Support, support@sponserity.com
