Kisop proposes a system to enable a privacy-focused personal health repository that allows AI agents to request consent for accessing health data while using personal foundation models. This approach keeps user data local and protected.
The system contemplates a privacy-first approach to personal health data management with the following key components:
- Personal Health Repository
- Securely stores all user health data locally
- Includes dedicated components for health records storage, consent management, and access control
- Owned and controlled entirely by the user
- Consent Management System
- Allows users to explicitly approve or deny access requests
- Manages granular permissions for different data types and uses
- Creates an audit trail of all access requests and decisions
- External AI Agents
- Must request specific access to user data
- Cannot access data without explicit user consent
- Receive only the minimum necessary data approved by the user
- Personal Foundation Models
- Run locally on the user’s secure environment (or in a Cloud environment)
- Provide AI capabilities without sending data to external systems
- Can be personalized based on the user’s health profile while preserving privacy
- No Spreading of Data throughout many systems
- Raw health data never leaves the user’s secure environment
- External AI models cannot train on the user’s data without permission
- Health Data Providers
- Can securely transfer data to the personal repository
- Data remains encrypted during transfer and at rest
- Once imported, data is controlled by the user’s consent system
This architecture ensures that users maintain complete control over their health information while still benefiting from AI-powered health insights. The personal foundation model allows for personalized AI assistance without compromising privacy, and the consent management system gives users transparency and choice in how their data is used.
Kisop allows end users to create an augmented personal repository that is portable, multilingual, extensible, and can interact with artificial intelligence systems while preserving privacy and security in a global system of cooperation.