MyOrbit Safety Policy

1. Our Safety Framework

1.1 Core Safety Commitments

We maintain zero tolerance for:

• Child sexual abuse material (CSAM) or child exploitation
• Terrorism content or promotion
• Non-consensual intimate content
• Self-harm instruction or promotion
• Severe harassment or doxxing
• Criminal activity coordination

1.2 MyOrbit Content Levels

MyOrbit sorts all content into four levels. Every account starts at General, and the two adult levels are opt-in only:

• General (all users, 13+; 16+ in the EU): The default for every account. Everyday conversation, education, creativity, business and support content, and sensitive topics such as mental health handled with added context and safety resources.
• Mature (18+, opt-in): Adult themes, adult humor, dating and relationship topics, entertainment violence, and alcohol or cannabis references where legal. Available only to accounts confirmed as adult, and only after the account holder turns it on in Settings.
• Adults Only (18+, opt-in): A separate, clearly labeled category for romantic and intimate storytelling. It requires the same confirmed-adult account as Mature plus its own opt-in in Settings; it is not a stricter identity check, only a separate choice. Adults Only content is non-explicit: intimate scenes fade to black, in text, voice and media alike. MyOrbit does not host sexually explicit media.
• Prohibited (blocked for everyone): Everything in our zero-tolerance list above, anything illegal, and anything that breaches our Terms of Service. No age, setting or account type unlocks it.

Minors aged 13-17, and any account linked to a parent through Parental Controls, are restricted to General content at all times; the Mature and Adults Only opt-ins are not offered to them. Adults Only content is withheld from any account that has not opted in: it is not served to them in Discover, in recommendations, or in the feed. Reaching it is a deliberate act, and the adult and opt-in check runs again before an Adults Only avatar can be downloaded.

1.3 Encryption and Safety

MyOrbit messaging runs in a single mode. Your messages are stored, not discarded after delivery. Traffic is encrypted in transit with TLS 1.3, and our databases, file storage and backups are encrypted at rest with AES-256 at the infrastructure layer, with additional application-level encryption on some message fields but not all of them. Message content passes through our safety systems so that the protections in this policy can work:

• Automated content scanning for safety
• Proactive threat detection
• Real-time intervention capabilities
• Crisis support features active
• Full moderation toolkit available

We do not offer a zero-access messaging mode. Please read this policy and our Privacy Policy on the understanding that MyOrbit is technically able to access message content when our safety, legal or support processes require it.

2. Content Moderation System

2.1 Layered Defense

What is actually in place:

• AI Moderation: OpenAI's multimodal moderation model screens message text and uploaded images (avatars, stories, chat attachments), plus our own safety layers in the prompt and response path
• API Gateway: Rate limiting and anomaly detection
• Human Review: a moderation team for nuanced decisions and escalations
• User Reporting: in-product reporting on profiles, messages, and content
• Post-Publication Monitoring: ongoing safety checks

We have removed two claims that were previously listed here: device-level client-side ML filtering, and CDN-level blocking of harmful content. Neither exists in the shipped product. Moderation happens server-side.

2.2 Reporting and Response

How to Report:

• Use in-app report buttons on any content
• Email safety@myorbit.ai for urgent issues
• Contact law enforcement directly for immediate danger

Our Response Timeline:

• Immediate: CSAM, terrorism, imminent harm
• Within 1 hour: Severe violations
• Within 6 hours: High-priority reports
• Within 24 hours: Standard reports
• Within 48 hours: Low-priority issues

2.3 Enforcement Actions

Based on severity and history:

• Warning: First-time minor violations
• Content Removal: Policy-violating material
• Feature Restrictions: Limited functionality
• Temporary Suspension: 24 hours to 30 days
• Permanent Ban: Severe or repeated violations
• Law Enforcement Referral: Illegal activities

2.4 Appeals Process

• Submit appeals within 30 days of action
• Different reviewer examines each appeal
• Response within 7 business days
• Final decisions may be escalated to senior team

3. AI Safety and Ethics

3.1 Transparent AI Interactions

All AI features include:

• Clear labeling as artificial intelligence
• Periodic reminders of AI nature
• Disclaimers about limitations
• Prohibition on impersonation
• Model information when relevant

3.2 AI Companion Boundaries

3.3 Preventing Unhealthy Dependencies

We monitor for and address:

• Excessive usage patterns
• Emotional dependency indicators
• Reality dissociation signs
• Isolation behaviors
• Crisis language

Interventions include:

• Usage reminders and breaks
• Reality check prompts
• Resource suggestions
• Professional help referrals
• Conversation limits when needed

3.4 Model Safety and Selection

• Different AI models for different content levels
• Automatic safety filtering on all outputs
• Refusal of harmful requests
• Bias monitoring and correction
• Regular safety audits

4. Child and Minor Safety

4.1 How We Establish Age

Every account gives a phone number at signup. To unlock the adult content levels described in section 1.2, an account must additionally be confirmed as adult by a selfie-based age estimate: you take a selfie in the app, it is used to check whether you are an adult, and it is not retained afterwards. This is an age estimate, not facial recognition, and it is not used to identify you.

If the estimate reads an adult as a minor, there is an appeal: you may choose to verify with a government ID through our identity partner, Stripe Identity. This route exists solely to correct a wrong estimate. Government ID is never a precondition for any content level, and there is no content on MyOrbit that requires it.

4.2 Minor-Specific Protections (13-17)

• Private profiles by default
• General content only - no Mature or Adults Only access, ever
• The Mature and Adults Only opt-ins are not offered
• Limited adult messaging
• Enhanced content filtering
• Parental control options
• Educational safety resources
• Restricted monetization features

4.3 Parental Controls

Parents/guardians can:

• Link to minor's account
• Monitor activity (with minor's knowledge)
• Set daily usage time limits through your device's built-in screen-time controls
• Review connections
• Request data deletion
• Receive safety alerts

4.4 COPPA Compliance (Under 13)

MyOrbit is not intended for children under 13, and accounts may not be created for them. In compliance with COPPA:

• We do not knowingly collect data from children under 13
• Accounts suspected of being under 13 are suspended and closed
• Any data from such accounts is deleted promptly
• Parents can contact privacy@myorbit.ai to request review or deletion

5. Creator and Business Safety

5.1 Creator Protections

• Identity verification for monetization
• "Look Ma I'm Famous" IP protection system
• 30-day grace period for unauthorized avatars
• Boxing system for celebrity likeness
• DMCA tools and processes
• Revenue protection measures

5.2 Business Twin Safeguards

• Verified business accounts only
• Accuracy disclaimers required
• Audit trails for all interactions
• Hallucination detection systems
• Human approval for high-stakes decisions
• Professional boundary enforcement

5.3 Community Standards for Creators

Creators must:

• Accurately label content levels
• Verify age-appropriate audiences
• Respect IP rights
• Maintain professional boundaries
• Follow monetization guidelines
• Respond to safety reports

6. Platform-Specific Safety Measures

6.1 Messaging Safety

• Spam and phishing detection
• Malware scanning
• Unsolicited content filtering
• Block and mute tools
• Encryption in transit and at rest
• Forward limiting

6.2 Group and Community Safety

• Admin moderation tools
• Community guidelines enforcement
• Member verification options
• Content filtering by age
• Reporting escalation paths
• Automatic content-level restrictions

6.3 Marketplace Safety

• Seller verification requirements
• Prohibited items enforcement
• Transaction monitoring
• Dispute resolution process
• Fraud detection systems
• INFORM Act compliance

7. Privacy-Preserving Safety

7.1 How We Balance Privacy and Safety

Safety work is scoped as narrowly as we can make it while still being effective:

• Automated scanning for severe violations
• Metadata analysis for patterns
• Crisis detection active
• User reporting with evidence
• Human review limited to flagged material

7.2 Data Minimization

We collect only what's necessary for safety:

• Basic account information
• Safety-relevant metadata
• User reports and evidence
• Age-check selfies, which are used for our age check and are not retained
• Appeal documents submitted to our identity partner (deleted after the appeal is decided)
• Behavioral patterns (anonymized)

8. Crisis Response and Support

8.1 Crisis Detection (AI Mode Only)

Automated detection for:

• Suicide or self-harm expressions
• Violence planning
• Child safety threats
• Medical emergencies
• Mental health crises

8.2 Crisis Response Protocol

• Immediate safety check
• Warm handoff to human support
• Crisis resources provided
• Professional referrals offered
• Follow-up when appropriate
• Never punitive for help-seeking

8.3 Emergency Resources

9. Legal Compliance and Cooperation

9.1 Regulatory Compliance

We comply with:

• GDPR: EU data protection and privacy
• DSA: Digital Services Act requirements
• AI Act: EU AI regulations (February 2025)
• COPPA: Children's privacy protection
• CCPA/CPRA: California privacy rights
• Section 230: Platform protections
• DMCA: Copyright enforcement

9.2 Law Enforcement Cooperation

What we can provide:

• Account information with valid legal process
• Available message content with warrant
• Emergency disclosure for imminent harm
• Preservation requests honored

We respond to valid legal process for the data we hold, and we require valid legal process before disclosing it.

9.3 Transparency Reporting

Quarterly reports include:

• Content removal statistics
• Account action metrics
• Law enforcement requests
• Appeal outcomes
• Safety investment updates

10. Safety Tools and Features

10.1 User Control Tools

• Block: Prevent all interaction
• Mute: Hide content without blocking
• Report: Flag violations for review
• Filter: Customize content visibility
• Restrict: Limit who can contact you
• Safety Mode: Enhanced protection settings

10.2 Proactive Safety Features

• Suspicious pattern detection
• New device login alerts
• Unusual activity warnings
• Password breach monitoring
• Two-factor authentication

10.3 Educational Resources

• Digital literacy guides
• Parent resources
• Creator safety handbook
• Business compliance guide
• Mental health resources
• Privacy best practices

11. Incident Response Plan

11.1 Security Incidents

• Within 1 hour: Containment and initial assessment
• Within 6 hours: Senior team notification
• Within 24 hours: User notifications if affected
• Within 72 hours: Regulatory notifications (GDPR)
• Within 7 days: Public transparency update
• Within 30 days: Complete post-mortem

11.2 Safety Incidents

• Immediate: Remove illegal content, secure evidence
• Within 1 hour: Escalate to safety team
• Within 6 hours: Law enforcement contact if required
• Within 24 hours: Affected user support
• Within 48 hours: Community update if needed

11.3 Post-Incident Actions

• Root cause analysis
• Process improvements
• Tool enhancements
• Training updates
• Policy revisions
• Community communication

12. Community Guidelines

12.1 Expected Behavior

• Treat others with respect
• Protect minors from harm
• Respect privacy and consent
• Share responsibly
• Report violations
• Support community safety

12.2 Prohibited Behavior

See Section 5.2 of our Terms of Service for complete list.

Key prohibitions:

• Illegal activities
• Harm to minors
• Non-consensual content
• Severe harassment
• Impersonation
• Platform manipulation

12.3 Cultural Sensitivity

We respect diverse communities while maintaining safety standards:

• Context considered in moderation
• Cultural expression protected
• Hate speech never tolerated
• Educational content preserved
• Artistic expression supported
• Religious freedom respected

13. Safety Partnerships

13.1 Industry Collaboration

We work with:

• National Center for Missing & Exploited Children (NCMEC)
• Technology Coalition
• Family Online Safety Institute (FOSI)
• Cyber Civil Rights Initiative
• Crisis Text Line
• Academic safety researchers

13.2 Technology Partners

• OpenAI: automated content moderation of message text and uploaded images, including blocking sexual content involving minors
• Anthropic: Safety alignment
• Age estimation and Stripe Identity: Adult confirmation and appeals

We do not use PhotoDNA, Perspective API, or WebPurify. Hash-matching against known-CSAM databases is in progress and is not live yet — see our Child Safety Policy for the current status.

14. Updates and Changes

14.1 Policy Updates

• Regular reviews (quarterly minimum)
• User notification of material changes
• 30-day notice before implementation
• Archived versions available
• Feedback incorporation

14.2 Continuous Improvement

We continuously enhance safety through:

• User feedback integration
• Incident learning
• Research partnerships
• Technology advancement
• Regulatory alignment
• Community input

15. Contact and Support

15.1 Safety Team Contact

15.2 Escalation Path

1. In-app reporting
2. Safety team email
3. Senior safety team escalation
4. Executive review (critical issues)
5. Board oversight (systemic concerns)

15.3 Office Hours

• Safety team: 24/7 coverage
• Urgent response: Always available
• Standard support: 24-48 hours
• Appeals: 7 business days

16. Accessibility and Inclusion

16.1 Inclusive Safety

Safety tools available for all users:

• Screen reader compatible reporting
• Multiple language support
• Clear, simple language options
• Visual and audio alternatives
• Assistive technology support

16.2 Vulnerable User Protection

Enhanced protections for:

• Users with disabilities
• LGBTQIA+ community members
• Minority communities
• At-risk populations
• Elderly users
• New internet users