Mental Health Context Protection

Graduated disclosure controls for sensitive mental health information.

Sensitive Context Settings

Configure how your mental health context is shared and used

Context Flags

These are shared as boolean values only (true/false)

Seeking Support
Professional Involved
Crisis Indicators
Escalation Consent

Requested Adaptations

How AI should adjust its behavior

Private Context (Never Transmitted)

You have private context configured. This information shapes AI behavior but is never shared with any stakeholder, even with "full" sharing enabled.

Conditions: 2 Triggers: 2 Coping: 2

What Stakeholders See

AI Assistant
Flags + adaptation preferences (3 adaptations active)
Human Stakeholders
Level: minimal Escalation consent: Yes

Mental Health Scenarios

Context Audit

5 shared 2 influenced 2 withheld
Shared 5
AI Sharing Level
moderate
User-controlled disclosure level for AI systems
Human Sharing Level
minimal
User-controlled disclosure level for human stakeholders
Seeking Support
Yes
Professional Involved
Yes
Escalation Consent
granted
User controls whether crisis escalation can occur
Influenced 2
Active Adaptations
3 adaptations
Influences AI communication style without exposing why
Sensitive Topics
1 topics marked
AI avoids or approaches carefully without knowing specifics
Withheld 2
Private Context
defined
Conditions, triggers, coping strategies NEVER transmitted
Crisis Indicators
inactive
Protected when inactive
Private context influenced behavior without being exposed

Privacy Guarantees

Private Context Never Transmitted

Conditions, triggers, and coping strategies in private_context are NEVER sent to any stakeholder, regardless of sharing level.

Minimal = Booleans Only

At minimal sharing, only yes/no flags are transmitted. No details about what kind of support, what conditions, or what adaptations.

AI vs Human Separation

You can share more with AI (which has no persistent memory) than with human stakeholders who might form lasting judgments.

Crisis Escalation Consent

Even if crisis indicators are detected, escalation only happens if escalation_consent is true. User maintains control.

How Adaptations Work

When you enable adaptations, AI systems adjust their behavior:

Gentle Language
"You failed to complete the task" → "This one didn't work out - let's try a different approach"
Avoid Pressure
"You need to finish this TODAY" → "When you're ready, we can work on this together"
Celebrate Small Wins
"Task completed." → "Great job getting that done! Every step forward counts."
Key: These adaptations happen automatically when your VCP context includes mental health preferences. You don't have to disclose WHY you need gentle language - the AI just provides it.