Content Moderation Policy

This Content Moderation Policy explains how OCAI LTD ("HoneyPot," "we," "our," or "us") moderates content on honeypot.com: what we look for, how we detect it, and how we act.

1. Why We Moderate

HoneyPot provides adult AI companions for users 18+. To keep the platform safe and lawful we apply continuous moderation to:

• Prevent generation of content prohibited by the Prohibited Content Policy
• Detect and block illegal content (CSAM, non-consensual intimate imagery, terrorism content, etc.)
• Enforce our Terms of Service
• Respond to reports from users, third parties, and authorities
• Comply with payment-processor, hosting-provider, and legal requirements

2. What We Moderate

Moderation applies to:

• User prompts: text sent to a Honey
• Honey configurations: names, descriptions, personality prompts, gallery uploads
• AI generations: text, image, video, and audio output produced by Honeys
• Profile information: usernames, bios, avatars
• Reports and support communications

3. Layered Approach

We combine multiple moderation layers; together they catch a wide range of violations while keeping latency low.

Layer 1: Pre-generation classifiers. Every user prompt and every Honey configuration is passed through our proprietary classifier stack before it reaches a generation model. The classifier scores against prohibited-content categories (minors, real people, hate speech, violence, illegal activity). High-confidence violations are blocked instantly.

Layer 2: In-model safety. Our underlying generation models include refusal training and safety RLHF that further reduce the rate of prohibited output for borderline prompts that pass Layer 1.

Layer 3: Post-generation review. Generated content is re-scanned by classifiers and hash-matched against known illegal-content databases (CSAM hashes via NCMEC's PhotoDNA, terrorist-content hashes via GIFCT, etc.). High-confidence matches are quarantined immediately.

Layer 4: Human review. Trained human moderators review:

• Content flagged by Layers 1-3 with intermediate confidence
• User and third-party reports
• Appeals from users whose content was actioned
• Random samples for quality assurance

Human reviewers operate under written guidelines aligned with this Policy and our Prohibited Content Policy. They have access to context (Honey configuration, conversation history) sufficient to make accurate calls.

Layer 5: Independent audit. We periodically engage independent reviewers to audit the moderation pipeline, sample our outputs, and verify that our enforcement matches our policies.

4. Actions We Take

Depending on severity, we may:

• Block the prompt or generation in real time and inform the user
• Remove the content and notify the user with the reason
• Temporarily suspend the account pending review
• Permanently terminate the account
• Revoke access to specific features (e.g. community publishing)
• Report to law enforcement and child-safety hotlines for content involving minors or imminent harm
• Preserve evidence and metadata where mandatory reporting requires

5. Reports and Appeals

Reports. Anyone can report content via support@honeypot.com or via the in-product report flow. We acknowledge reports within 5 business days and prioritize categories defined in the Content Removal Policy.

Appeals. Users whose content was actioned may appeal as described in the Content Removal Policy. Appeals on expedited-removal categories are final.

6. Transparency

We publish summary statistics about enforcement actions on a regular cadence (the first public report is planned within the first 12 months of the Services going live). The report will cover:

• Volume of content moderated by category
• Volume of accounts actioned
• Volume of reports received and resolution times
• Volume of law-enforcement requests received and complied with

7. Privacy and Confidentiality

Moderation may involve human review of user content. Reviewers operate under confidentiality agreements and access logs are retained for audit. We don't use the content of personal conversations for marketing or for any purpose beyond moderation, security, and the operational uses described in our Privacy Policy.

Contact

support@honeypot.com

OCAI LTD
Nicosia, Cyprus