Anti-Harassment Filters Policy

1. Purpose & Safety Objectives

Friendium is designed to foster respectful, authentic interaction between real people. Anti-harassment filters exist to proactively reduce harm, minimize exposure to abusive content, and empower users to control their experience without requiring constant manual moderation.

These systems are a core component of Friendium’s safety architecture and complement Community Standards, Reporting mechanisms, and human review.

2. Scope of Filters

Anti-harassment filters apply across multiple interaction surfaces, including:

• Comments and replies
• Direct messages (DMs)
• Mentions and tags
• Profile posts and wall interactions
• Group discussions and event comments

3. Categories of Filtered Content

Filters may automatically detect, flag, limit, or block content associated with:

• Harassment, bullying, or intimidation
• Hate speech or discriminatory language
• Threats of violence or harm
• Sexual harassment or degrading remarks
• Repeated unwanted contact
• Obscene or aggressive language patterns

4. Keyword & Phrase Filtering

Friendium uses keyword-based filtering to identify commonly abusive or harmful terms. Users may also:

• Add custom keywords to personal filter lists
• Block phrases associated with harassment
• Filter content in multiple languages
• Apply filters selectively to comments or messages

5. Behavioral Pattern Detection

In addition to keywords, Friendium evaluates behavioral signals, including:

• High-frequency messaging to a single user
• Repeated negative interactions across posts
• Coordinated harassment patterns
• Escalating tone or language severity

These systems are designed to identify harassment even when explicit slurs are absent.

6. Automated Action Types

Depending on severity and context, filters may:

• Hide content from the recipient
• Require manual approval before display
• Limit visibility to the sender only
• Trigger warning screens
• Escalate content for human review

7. User-Controlled Filter Settings

Users can customize their anti-harassment protections by:

• Enabling strict, standard, or relaxed filtering modes
• Filtering content from non-connections
• Restricting messages from newly created accounts
• Blocking repeated commenters automatically

8. Protection for Vulnerable Users

Additional safeguards may be automatically enabled for:

• Minors and young adults
• Users experiencing active harassment campaigns
• Users targeted due to protected characteristics
• Users involved in sensitive or high-risk discussions

9. Transparency & Explainability

When content is filtered or limited, Friendium may provide:

• Notifications explaining the restriction
• General reasoning categories (e.g., harassment detection)
• Links to relevant policies

To prevent abuse of the system, exact detection thresholds are not disclosed.

10. False Positives & User Appeals

Friendium acknowledges that automated systems are not perfect. Users may:

• Appeal filtered or restricted content
• Request human review
• Provide contextual explanations

Successful appeals may improve future filter accuracy.

11. Interaction With Other Safety Systems

Anti-harassment filters work alongside:

• Blocking and muting tools
• Restricted lists
• Reporting & escalation workflows
• Repeat offender tracking

12. Enforcement Escalation

Repeated triggering of filters may lead to:

• Temporary messaging or commenting limits
• Account feature restrictions
• Mandatory safety acknowledgments
• Suspension or permanent removal

13. Good-Faith Expression

Filters are not intended to suppress respectful disagreement, satire, or critical discussion. Friendium balances safety with freedom of expression by evaluating intent, context, and harm.

14. Legal & Regulatory Alignment

Filtering systems are designed to support compliance with global online safety, anti-harassment, and child-protection laws, while respecting lawful speech rights.

15. Continuous Improvement

Friendium regularly updates filter models based on:

• User feedback
• Moderator review outcomes
• Emerging abuse patterns
• Regulatory guidance

16. Policy Updates

This policy may be revised as new risks, technologies, or legal obligations arise. Continued use of Friendium constitutes acceptance of updated filtering practices.

17. Contact & Support

Safety & Harassment Reports: safety@friendium.com
Policy Questions: trust@friendium.com
User Support: support@friendium.com