🛡️ Safety Design

AI Buddha Zen implements a multi-layered safety framework to protect users. This page explains our approach to crisis response, professional boundary setting, and privacy protection.

🛡️ Crisis Message Priority Response

🚨 The Problem We Solved NEW

Many AI chat services set daily usage limits on free plans. But what happens when someone sends "I want to die" after reaching their limit? Most services respond with "Daily limit reached." AI Buddha Zen was designed to never let this happen.

How it works:

1User sends a message
2Crisis keyword detection (BEFORE rate limit check)
3If crisis detected → bypass rate limit → respond normally + suggest helpline
4If 3+ crisis messages in 24h → stop AI response, show helplines only (escalation)

Why Escalation Matters

When the same user sends repeated crisis messages, it indicates the AI's words alone are insufficient. At that point, continuing to generate AI responses could delay the person from seeking real human support. The escalation feature recognizes this boundary and directly connects the user to professional helplines.

🩺 Two-Layer Response Design

🩺 For Medical, Legal, and Financial Concerns NEW

When a user asks about medication effects or legal disputes, AI Buddha Zen responds in two layers. First, it clearly states "Please consult a professional." Then, it offers to support the emotional aspect of the situation through Buddhist wisdom.

❌ What we avoid

"Just take a deep breath" — responding with only spiritual practice when the user needs medical help feels dismissive.

✓ What we do

"Please consult your doctor about medication. However, I can share Buddhist wisdom about the difficulty of waiting." — Clear boundary + emotional support.

🚨 C-SSRS 3-Tier Suicide Risk Detection

Based on the Columbia-Suicide Severity Rating Scale (Posner et al., 2011, American Journal of Psychiatry), suicide risk is classified into three tiers:

Tier 3 — Imminent Risk

Specific methods or plans mentioned. Immediate crisis intervention with hotline referral.

Tier 2 — Active Ideation

Explicit wish to die without specific plan. Crisis message with helpline information.

Tier 1 — Passive Ideation

Indirect expressions of hopelessness. AI adjusts response to empathetic mode and gently mentions professional resources.

🧘 Additional Safety Features

🧘Spiritual Bypassing Detection (SBS-13)

Based on Fox, Cashwell & Picciotto (2017), detects when users may be using spiritual concepts to avoid confronting psychological pain. The AI acknowledges real feelings rather than offering superficial spiritual comfort.

🔗AI Dependency Risk Monitoring

Detects signs of emotional dependency on AI through message content analysis and usage patterns. The AI gently encourages real human connections when dependency signals are detected.

🔐Cryptographic Privacy (CAP-SRP v2.0)

All risk assessments are recorded in a cryptographic audit chain (SHA-256 + HMAC). Original message text is never stored — only one-way hashes. An IETF Internet-Draft has been submitted for standardization.

If You Need Help Right Now

📞 0120-279-338
Yorisoi Hotline — 24 hours, free, anonymous
📞 0570-783-556
Inochi no Denwa (Lifeline)
📞 0570-064-556
Mental Health Consultation Dial
💬 Chat-based counseling (comarigoto.jp)

Important Disclaimer

AI Buddha Zen is an AI dialogue service based on Buddhist scriptures. It does not provide medical advice, diagnosis, treatment, counseling, or any other medical services. AI responses are provided as reference information and do not replace professional advice from physicians, psychiatrists, psychologists, or other healthcare professionals. The "Crisis Message Priority Response" and "Suicide Risk Detection" features are supplementary mechanisms designed to connect users with professional helplines and do not perform clinical judgments or diagnoses.

💬 Try on LINE App Store