Trusted Device
Important actions build on a trusted device and its current security state.
SECURITY & PRIVACY
Security built around trust, permission and user control.
MiiA does not treat “AI can do this” as “AI is allowed to do this.” Security and privacy boundaries come before assistance and automation.
DESIGN PRINCIPLES
Clear boundaries from the device to the relationship.
These principles describe MiiA’s product and architecture direction without relying on absolute or unverifiable security claims.
Platform Authentication
When re-authentication is required, iOS or Android owns the native security presentation.
Privacy Lock
Specific conversations can have an additional privacy and re-authentication boundary.
Access Stays Closed
When authorization is unclear, access and sharing scope do not expand.
Minimum Permissions
Request only the permissions actually needed for the task at hand.
Trusted Relationships
Trust comes from relationships the user explicitly establishes or verifies, not AI inference alone.
Protected Notifications
Sensitive content should remain appropriately redacted in notification and preview contexts.
AI Permission Boundary
AI assistance follows existing relationship, sharing and authorization boundaries.
User-Controlled Memory
Personal memory is designed around user control, confirmation and manageability.
AI PERMISSION BOUNDARY
Establish what is allowed before MiiA assists.
Permission and authentication are prerequisites MiiA must respect, not conclusions AI is allowed to invent.
Trust grants coordination; AI never manufactures authority.
- 01Relationship
- 02Permission
- 03Authentication
- 04Allowed Action
- 05MiiA Assistance
USER CONTROL
Security should not make the experience harder to understand.
MiiA security experiences should be clear and restrained so users understand when authentication is required, when content is protected and where the sharing boundary sits.
Clear next actions
When authentication or permission is needed, show an understandable next step instead of internal diagnostics.
No imitation system prompts
MiiA does not create fake Face ID, Touch ID or Android BiometricPrompt experiences.
No exaggerated security claims
Trust is not built with absolute security language that lacks validated evidence.
SECURITY REPORTING
Security reporting channel
The official security reporting method will be published after a real, tested MiiA contact channel is ready to receive reports. No unverified channel is presented as active.