Crisis Response Policy
1. Our Commitment to Crisis Response
Quibit, Inc. ("Quibit," "we," "us," or "our") is committed to protecting the safety and well-being of every user across all of our applications: Quibit (social), Kha (dating, 18+), QuiTalk (AI chat, channels, and phone calls), and QuizCoin (quiz game). When a crisis or emergency arises — whether involving imminent danger to life, exploitation, or threats of harm — we act swiftly and decisively.
This Crisis Response Policy outlines our protocols for identifying, responding to, and following up on crisis situations. It applies to all content, communications, AI-generated responses, phone calls, and user interactions across every Quibit, Inc. application.
Our crisis response framework operates 24/7 and is designed to work in coordination with law enforcement, emergency services, mental health professionals, and child safety organizations worldwide.
If you or someone you know is in immediate danger, call your local emergency services (911 in the U.S., 112 in Korea, 999 in Myanmar) before using any in-app reporting features. Quibit is not a substitute for emergency services.
2. Types of Crises Covered
This policy covers a broad range of crisis situations that may arise on or be communicated through Quibit platforms. Our response protocols are calibrated to the severity and nature of each crisis type.
2.1 Self-Harm and Suicide
Any expression of suicidal ideation, self-harm intent, or active self-injury communicated through posts, messages, AI conversations, or phone calls.
Direct statements of suicidal intent or planning
Self-harm imagery or descriptions
Suicide notes shared on the platform
Farewell messages suggesting imminent self-harm
AI conversations where users express hopelessness or desire to end their life
2.2 Domestic Violence and Intimate Partner Abuse
Reports or indicators of domestic violence, including physical, emotional, sexual, or financial abuse by a partner or family member.
Users disclosing abuse in messages or AI conversations
Reports of controlling or threatening behavior by a dating match (Kha)
Evidence of coercive control communicated through the platform
2.3 Child Sexual Exploitation and Abuse
Any indication of child sexual abuse material (CSAM), grooming, trafficking, or exploitation. See our Child Safety Standards Policy for comprehensive protocols.
CSAM detected on any platform
Grooming behavior in messages or AI conversations
Sextortion targeting minors
Trafficking indicators across any app
2.4 Active Threats and Violence
Credible threats of violence against individuals, groups, or locations communicated through any Quibit platform.
Threats of mass violence or terrorism
Targeted threats against specific individuals
Threats involving weapons or explosives
Incitement of real-world violence through posts, channels, or messages
2.5 Severe Cyberbullying and Harassment
Sustained, severe harassment campaigns that pose a credible risk to the target's safety or mental health.
Coordinated harassment campaigns
Doxxing (publishing private information)
Persistent threats or intimidation
Harassment escalating to real-world stalking or threats
2.6 Sextortion and Image-Based Abuse
Coercion or blackmail using intimate images, whether real or AI-generated, targeting users of any age.
Threats to distribute intimate images without consent
Demands for money, sexual content, or other actions under threat of exposure
Non-consensual sharing of intimate images
AI-generated deepfake intimate content used for harassment
2.7 Natural Disasters and Public Emergencies
Large-scale events that affect user safety and may require platform-level coordination.
Earthquakes, typhoons, floods, and other natural disasters
Public health emergencies and pandemics
Civil unrest or conflict in regions with active users
Infrastructure failures affecting user safety
2.8 Human Trafficking
Indicators of human trafficking, forced labor, or exploitation communicated through platform features.
Suspicious dating profiles or activity on Kha suggesting trafficking
Coded language or recruitment attempts in messages
Reports of forced labor or exploitation
3. Immediate Response Protocols
When a crisis is detected — whether by automated systems, AI monitoring, user reports, or staff review — Quibit follows a structured escalation protocol to ensure the fastest possible response.
3.1 Triage and Classification
All incoming crisis signals are classified by severity to determine response speed and escalation path:
| Violation Type | Enforcement Action |
|---|---|
| Tier 1 — Imminent Danger (Life Threat) | Immediate response within 15 minutes. Emergency services contacted. Content preserved. Examples: active suicide attempt, ongoing violence, child in immediate danger. |
| Tier 2 — Urgent (Potential Danger) | Response within 1 hour. Escalated to Trust & Safety lead. Examples: credible threats, sextortion in progress, domestic violence disclosure. |
| Tier 3 — Serious (Requires Intervention) | Response within 4 hours. Assigned to specialized team member. Examples: sustained harassment, trafficking indicators, severe cyberbullying. |
| Tier 4 — Monitoring (Elevated Concern) | Response within 24 hours. Logged and monitored for escalation. Examples: vague threats, concerning behavioral patterns, repeated concerning AI interactions. |
3.2 Immediate Actions Taken
Depending on the crisis tier, the following actions may be taken immediately:
Contact emergency services (911, 112, or local equivalent) for Tier 1 incidents when user location is known or can be determined
Preserve all relevant content, metadata, and conversation history for law enforcement
Restrict or suspend accounts involved in perpetrating harm
Display crisis intervention resources to the affected user (hotline numbers, safety tips)
Notify the Trust & Safety on-call team via internal alert system
Block, mute, or restrict the aggressor's ability to contact the victim
Temporarily restrict the affected user's ability to delete evidence (with legal basis)
3.3 Crisis Intervention Messages
When self-harm or suicide signals are detected, Quibit may display intervention messages directly in the app:
Information about the 988 Suicide & Crisis Lifeline (U.S.) and equivalent services in other countries
A prompt to text or call a crisis counselor
Links to safety planning resources
Option to connect with a trusted contact
Reminder that help is available and the user is not alone
Quibit may automatically surface crisis resources and helpline information when our systems detect crisis signals. This feature cannot be disabled, as it serves a life-safety function.
4. AI Crisis Detection (QuiTalk)
QuiTalk's AI chat system includes specialized crisis detection capabilities that monitor conversations for signals of distress, danger, or harm. These systems are designed to protect users while respecting privacy and providing appropriate support.
4.1 AI Conversation Monitoring
QuiTalk's AI models are trained to recognize crisis signals in real time during conversations:
Expressions of suicidal ideation, self-harm intent, or hopelessness
Descriptions of ongoing abuse, violence, or exploitation
Crisis keywords and phrases associated with imminent danger
Escalating emotional distress patterns across multiple messages
Requests for help with dangerous situations
Indicators of sextortion, blackmail, or coercion
4.2 Crisis Keyword Detection
Our systems monitor for crisis-related language patterns across all supported languages (English, Korean, Burmese, Thai, Japanese). Detection includes:
Direct crisis statements (e.g., 'I want to kill myself,' 'I am being abused')
Indirect crisis signals (e.g., 'I can't go on,' 'no one would miss me,' 'there's no way out')
Coded language and euphemisms commonly associated with crisis situations
Context-aware analysis that distinguishes genuine distress from casual conversation
Multilingual detection across all languages supported by QuiTalk
4.3 AI Safety Responses
When QuiTalk's AI detects a crisis signal, it responds with built-in safety protocols:
The AI acknowledges the user's distress with empathy and validation
The AI provides relevant crisis hotline numbers and resources based on the user's region
The AI encourages the user to reach out to a trusted person or professional
The AI does not provide instructions for self-harm, violence, or illegal activity under any circumstances
The AI avoids dismissive or minimizing language
The AI may suggest the user contact emergency services if the situation appears imminent
When crisis signals are detected, QuiTalk's AI safety protocols override all other conversation context, channel brain settings, and user preferences. The AI will always prioritize user safety over conversational continuity.
4.4 Automatic Escalation
Certain crisis signals trigger automatic escalation beyond the AI response:
Tier 1 crisis signals are immediately flagged to the Trust & Safety on-call team
Repeated or escalating crisis signals within a single conversation trigger human review
AI-detected CSAM or exploitation language triggers immediate account investigation
Crisis conversations are logged (with appropriate privacy protections) for safety review
Users in active crisis may be offered direct connection to a human crisis counselor when available
4.5 Phone Call Crisis Detection
QuiTalk's phone calling feature includes additional crisis safeguards:
Voice-based AI interactions include the same crisis detection protocols as text conversations
Users who express distress during phone calls receive verbal crisis resource information
Phone call metadata (duration, frequency) is monitored for patterns associated with crisis situations
Emergency call routing information is provided when distress is detected during a call
5. App-Specific Crisis Protocols
Each Quibit, Inc. application has tailored crisis response protocols that address the unique risks associated with its features and user base.
5.2 Kha (Dating, 18+) — Dating Safety and Stalking Prevention
Kha's dating features present unique crisis risks including stalking, harassment, and unsafe meetups:
Users can activate a 'Safety Check' feature before meeting a match in person — the app sends timed check-in prompts and alerts an emergency contact if no response is received
Reports of stalking or harassment by a match are treated as Tier 2 crises with immediate investigation
Users who report domestic violence or intimate partner abuse through Kha receive tailored safety resources including local shelter information
Location sharing for dates includes an emergency SOS feature that sends location to trusted contacts
Accounts reported for threatening behavior are immediately restricted from contacting the reporter
Pattern detection identifies users who exhibit controlling or obsessive messaging behavior
Date safety tips and consent education are displayed before first messages
5.3 QuiTalk (AI Chat, Channels, Phone Calls) — Conversation Crisis Response
QuiTalk handles crisis situations across AI conversations, user-to-user channels, and phone calls:
AI conversations: Crisis detection and safety responses as described in the AI Crisis Detection section
Channels: Channel owners and moderators are trained to recognize and escalate crisis situations
Channel content promoting self-harm, violence, or exploitation is immediately removed
Phone calls: Voice-based crisis detection with automatic resource provision
Users in crisis during phone calls receive on-screen crisis hotline information
Channel AI brain content is reviewed to ensure it does not encourage harmful behavior
Emergency contacts can be pre-configured for automatic notification during detected crises
5.4 QuizCoin (Quiz Game) — Minor Safety and In-Game Crisis
QuizCoin's game environment requires special attention to minor safety and in-game communication:
In-game chat is monitored for bullying, threats, and crisis language
Minor accounts have enhanced chat filtering and crisis keyword detection
Users exhibiting distress in game chat receive automated crisis resource messages
Reports of predatory behavior toward minors in QuizCoin trigger Tier 1 escalation
Game-related frustration and anger are distinguished from genuine crisis signals through contextual analysis
Parental notifications are sent when minor accounts trigger crisis detection systems
6. Emergency Contacts and Resources
Quibit maintains a comprehensive directory of emergency contacts and crisis resources. These resources are surfaced to users in crisis based on their detected or self-reported location.
If you or someone you know is in immediate physical danger, always contact local emergency services (911, 112, 119, 999) before using in-app reporting. Quibit's crisis response supplements — but does not replace — professional emergency services.
7. Law Enforcement Cooperation
Quibit cooperates fully with law enforcement agencies worldwide to protect user safety. During crisis situations, our cooperation follows expedited procedures designed to minimize response time.
7.1 Emergency Data Preservation
When a crisis involves potential criminal activity or imminent danger, Quibit preserves relevant data for law enforcement:
Account information (registration data, profile details, device identifiers)
Content and communication records relevant to the crisis
IP addresses, login timestamps, and geolocation data where available
AI conversation logs related to the crisis
Phone call metadata and relevant records
Channel content and moderation history
Data is preserved for a minimum of 90 days pending law enforcement request, or longer if required by applicable law.
7.2 Emergency Disclosure Without Court Order
Under U.S. law (18 U.S.C. section 2702(b)(8)) and equivalent provisions in other jurisdictions, Quibit may disclose user information to law enforcement without a court order when we reasonably believe that an emergency involving danger of death or serious physical injury requires disclosure without delay. This includes:
Active suicide attempts or imminent self-harm
Ongoing violence or threats of imminent violence
Child exploitation in progress
Kidnapping or human trafficking situations
Active shooter or mass violence threats
7.3 Legal Process for Non-Emergency Requests
For non-emergency law enforcement requests, we require appropriate legal process:
Subpoenas for basic subscriber information
Court orders for transactional records and metadata
Search warrants for content of communications
Mutual Legal Assistance Treaty (MLAT) requests for international cases
Preservation requests to hold data pending legal process
Law enforcement should direct all requests to [email protected] with the subject line 'Law Enforcement Request — [Agency Name]'.
8. User Safety Tools
Quibit provides users with a range of safety tools designed to help them protect themselves and others during crisis situations.
8.1 Blocking and Reporting
All Quibit apps provide comprehensive blocking and reporting tools:
One-tap blocking that immediately prevents all contact from the blocked user across all Quibit apps
Dedicated crisis report categories (Self-Harm, Violence, Exploitation, Harassment) that trigger priority review
Anonymous reporting that protects the reporter's identity
Bulk reporting for coordinated harassment or abuse campaigns
Ability to report content, accounts, messages, AI conversations, and channel content
8.2 SOS and Emergency Features
Emergency features available across Quibit apps:
SOS button in settings that immediately surfaces local emergency numbers and crisis hotlines
Emergency contact designation — users can pre-select trusted contacts who are notified during crisis detection
Safety Check feature on Kha for in-person meetups with timed check-ins
Quick-exit feature that immediately closes the app and clears recent activity from the screen (for users in abusive situations)
Shake-to-report gesture for discreet crisis reporting when typing is not safe
8.3 Privacy and Safety Controls
Users can configure privacy settings to enhance their safety:
Control who can message, call, or view their profile
Hide online status and last-seen information
Disable location sharing and remove location from posts
Restrict comments and interactions on posts
Limit discoverability in search and recommendations
Two-factor authentication to prevent unauthorized account access
9. Mental Health Support
Quibit recognizes the critical connection between platform use and mental health. We are committed to providing support resources and safeguarding users who may be experiencing mental health crises.
9.1 In-App Mental Health Resources
Quibit provides mental health resources accessible from within all apps:
Mental health resource hub with articles, self-help tools, and professional directory
Guided breathing and grounding exercises accessible from the safety menu
Links to free and low-cost counseling services by region
Well-being check-in prompts for users who show signs of distress over time
Content sensitivity warnings on posts discussing mental health topics
9.2 AI Conversation Safeguards (QuiTalk)
QuiTalk's AI is specifically designed with mental health safeguards:
The AI does not act as a therapist or provide clinical mental health advice
The AI clearly directs users to professional resources when mental health concerns arise
The AI avoids reinforcing negative thought patterns or catastrophizing
Conversations about self-harm are handled with empathy and immediate resource provision
The AI does not provide information on methods of self-harm or suicide under any circumstances
Channel AI brains are prevented from generating content that could trigger or worsen mental health crises
Extended conversations about distress trigger automatic well-being check-in prompts
9.3 Content Safeguards
Platform-wide content policies support mental health:
Content promoting or glorifying self-harm, eating disorders, or suicide is prohibited
Recovery and support content is allowed and encouraged with appropriate sensitivity labels
Graphic descriptions of self-harm methods are removed regardless of context
Algorithms are tuned to avoid recommending distressing content to vulnerable users
Content warnings are applied to discussions of sensitive mental health topics
10. Post-Crisis Follow-Up
Quibit's responsibility does not end when the immediate crisis is resolved. We provide follow-up support to affected users and take steps to prevent recurrence.
10.1 Account Recovery and Restoration
For users whose accounts were affected during a crisis response:
Accounts restricted for protection during a crisis are reviewed for restoration within 48 hours
Users can appeal account actions taken during crisis response through a dedicated review process
Content removed during a crisis is preserved and may be restored if appropriate
Users affected by harassment or abuse receive guidance on securing their accounts
Password resets and security reviews are offered to users whose accounts may have been compromised
10.2 Ongoing Support Resources
After a crisis is resolved, Quibit provides continued support:
Follow-up messages with relevant support resources sent 24 hours and 7 days after a crisis
Warm referrals to professional mental health services based on user's location
Continued monitoring for signs of recurring distress (with user consent)
Option to connect with peer support communities
Safety plan creation tools and resources
10.3 Prevention and System Improvement
Each crisis informs our ongoing improvement efforts:
All crisis responses are reviewed in post-incident analysis
Detection algorithms are updated based on new patterns and false negatives
Response time metrics are tracked and optimized
User feedback on crisis response is solicited and incorporated
Quarterly crisis response audits assess protocol effectiveness
11. Staff Training and Preparedness
Quibit invests in comprehensive training to ensure our team is prepared to respond effectively to any crisis situation.
11.1 Trust & Safety Team Training
Our dedicated Trust & Safety team receives specialized training:
Initial 40-hour crisis response certification program
Monthly scenario-based drills covering all crisis types
Training on trauma-informed communication and de-escalation
Cultural competency training for our multilingual user base (English, Korean, Burmese, Thai, Japanese)
Legal compliance training on evidence preservation and law enforcement cooperation
Annual recertification in crisis intervention techniques
11.2 All-Staff Awareness Training
Every Quibit, Inc. employee receives baseline crisis awareness training:
Onboarding module on recognizing crisis indicators and escalation procedures
Annual refresher training on crisis response policies
Clear internal escalation procedures — every employee knows how to reach Trust & Safety immediately
Vicarious trauma support and mental health resources for staff who encounter crisis content
11.3 External Expert Partnerships
Quibit partners with external organizations to enhance our crisis response capabilities:
Consultation with suicide prevention experts on AI safety responses
Partnership with domestic violence organizations for resource referrals
Collaboration with NCMEC and law enforcement on child exploitation response
Engagement with mental health professionals for protocol review
Participation in industry working groups on platform safety and crisis response
12. Contact Information
For questions, concerns, or to report a crisis situation, contact Quibit through the following channels:
For related policies, please also review our Child Safety Standards Policy, Community Guidelines, and Privacy Policy. Together, these documents form Quibit's comprehensive approach to user safety across all applications.
For questions about this policy, contact [email protected]
© 2026 Quibit, Inc. · Version 1.0· Last updated 2026-03-13