AI Safety & Model Documentation

Last updated: 5/17/2026

We are committed to providing safe, responsible AI coaching experiences. This page outlines our AI models, safety measures, and content moderation policies.

AI Models & Technology

Our platform utilizes industry-leading large language models (LLMs) from trusted providers:

Primary Model: We use Google Gemini, selected for its strong safety track records and built-in content filters
Voice Technology: We integrate with Google Cloud for natural voice synthesis and optional voice cloning, which includes voice verification and consent mechanisms
Regular Updates: We stay current with the latest model versions to benefit from ongoing safety improvements

Content Moderation & Safety Filters

All AI-generated content passes through multiple layers of safety filters:

Built-in Model Safety: Our AI providers implement comprehensive safety filters that detect and block harmful content including hate speech, violence, sexual content, and discriminatory language
Input Filtering: User inputs are screened for inappropriate content before processing
Output Filtering: AI responses are evaluated for safety before being delivered to users
Conversation Monitoring: We maintain audit logs of conversations to identify and address potential safety issues
Creator Controls: Creators can configure additional guardrails and content boundaries for their specific AI agents

Mental Health & Crisis Limitations

Our AI agents are not substitutes for mental health professionals or crisis intervention services. Important limitations and protocols:

Not Crisis Intervention Tools: Our AI agents are not designed to detect or respond to mental health crises, suicidal ideation, or self-harm. They should never be used as a substitute for professional mental health services or crisis intervention
Crisis Resources: We strongly recommend that creators working in sensitive contexts (such as mental health, EAP, or counseling) configure their agents with prominent disclaimers and crisis resource information (e.g., National Suicide Prevention Lifeline: 988, Crisis Text Line: text HOME to 741741)
Medical Advice Limitations: AI agents are configured to state they cannot provide medical, psychiatric, or emergency health advice and should direct users to qualified professionals
Creator Responsibility: Creators are responsible for ensuring their AI agents include appropriate disclaimers, limitations, and resource links for their specific use case

Prohibited Content & Discrimination

We maintain a zero-tolerance policy for harmful content:

Racism & Discrimination: Our AI models are designed to refuse generating content that promotes racism, discrimination, or bias based on race, ethnicity, religion, gender, sexual orientation, disability, or any protected characteristic
Hate Speech: Content promoting hatred, violence, or dehumanization of individuals or groups is automatically blocked
Harassment: AI agents will not participate in or facilitate harassment, bullying, or targeted attacks
Illegal Activities: Content promoting illegal activities is prohibited and filtered
Misinformation: We implement measures to reduce the generation of false or misleading information, particularly on sensitive topics

Child Safety & Age Restrictions

Protecting children is a priority:

Account Requirements: Our platform requires account creation to access AI agents. We do not knowingly collect personal information from children under 13 in accordance with COPPA
Creator Responsibility: Creators who make their AI agents publicly accessible should ensure age-appropriate content and configure appropriate disclaimers
Child-Safe Content: Our AI models include built-in safety filters that help prevent inappropriate content
No Targeting of Minors: AI agents cannot be configured to specifically target or collect information from children
Parental Supervision: We recommend parental supervision for users under 18
Educational Use: For educational contexts, we recommend supervised use and age-appropriate content configuration

AI Limitations & Disclaimers

It is important to understand the limitations of AI technology:

Not Professional Services: AI agents are educational and informational tools, not replacements for licensed professionals in healthcare, legal, financial, or therapeutic contexts
Accuracy: While trained on creator content, AI agents may occasionally produce inaccurate or incomplete information
Bias: Despite safety measures, AI models may reflect biases present in their training data. We actively work to identify and mitigate these
Context Understanding: AI may misinterpret nuanced or ambiguous inputs
No Guarantee: We cannot guarantee AI responses will always be appropriate, accurate, or suitable for every situation

Accountability & Reporting

We maintain accountability through transparent processes:

Report Harmful Content: Users can report inappropriate AI responses or safety concerns through our platform
Creator Responsibility: Creators are responsible for the training content they provide and the guardrails they configure
Platform Oversight: We review reported content and may suspend or remove AI agents that violate our safety policies
Continuous Improvement: We regularly review safety incidents to improve our filters and protocols
Transparency: We maintain documentation of our safety measures and update this page as our practices evolve

Compliance & Standards

Our AI safety practices align with industry standards:

COPPA Compliance: We comply with the Children's Online Privacy Protection Act by not knowingly collecting personal information from children under 13
GDPR Considerations: Our data handling practices respect user privacy rights
Industry Best Practices: We follow guidance from AI safety research organizations and regulatory bodies
Model Provider Policies: We adhere to the acceptable use policies of Google Gemini
Regular Audits: We conduct internal reviews of safety systems and update protocols based on emerging best practices

Enterprise & EAP Deployment

For organizations deploying our AI coaches in sensitive contexts such as Employee Assistance Programs:

Enhanced Safety Configuration: We work with enterprise clients to configure additional safety guardrails specific to their context
Custom Disclaimers: Organizations can add custom disclaimers and resource links appropriate to their employee base
Audit Logs: Comprehensive conversation logs (privacy-respecting) for compliance and quality assurance
Professional Integration: AI coaches complement, not replace, human professionals. We support workflows that escalate to qualified staff
Contractual Commitments: Enterprise agreements can include specific safety SLAs and incident response protocols

Updates & Contact

AI safety is an evolving field. We continuously update our practices as technology and best practices advance. For questions about our AI safety measures or to report concerns, please contact us at our contact email.