Skip to main content

Understanding Risk Scores

Learn how Cirvia Parental's AI system assesses content and assigns risk scores to help you prioritize your child's online safety.

Risk Score Scale​

🟢 Low Risk (Scores 1-4)​

Normal childhood online activity

What it means:

  • Typical conversations and interactions
  • Age-appropriate content and language
  • General monitoring for awareness

Common examples:

  • "Good game everyone!"
  • "What's your favorite Roblox game?"
  • Sharing memes or funny images
  • Normal competitive gaming chat

Your action: Content is logged for your review but doesn't require immediate attention.

🟡 Medium Risk (Scores 5-7)​

Content that warrants closer attention

What it means:

  • Potentially concerning language or behavior
  • Social conflicts that might need guidance
  • Content that could escalate if not addressed

Common examples:

  • "Nobody likes you anyway"
  • Sharing personal information (school name, age)
  • Mild inappropriate language
  • Arguments with other users

Your action: Review content within 24 hours and consider talking with your child about appropriate online behavior.

🔴 High Risk (Scores 8-10)​

Serious safety concerns requiring immediate attention

What it means:

  • Direct threats to your child's safety
  • Predatory behavior from adults
  • Serious cyberbullying or harassment
  • Content that may require law enforcement involvement

Common examples:

  • "Want to meet up in person?"
  • Requests for photos or personal information
  • Sexual content or advances
  • Threats of violence or self-harm

Your action: Investigate immediately, document evidence, and take appropriate protective measures.

AI Analysis Categories​

Harassment Detection​

What we look for:

  • Bullying and intimidation tactics
  • Repeated negative targeting
  • Exclusion and social manipulation
  • Threats and aggressive language

Risk factors that increase score:

  • Personal attacks about appearance, abilities, or identity
  • Coordinated group harassment
  • Threats of physical harm
  • Doxxing or sharing personal information maliciously

Sexual Content Detection​

What we look for:

  • Inappropriate sexual language or requests
  • Sharing or requesting intimate images
  • Sexual predation tactics
  • Age-inappropriate romantic advances

Risk factors that increase score:

  • Adult requesting private communication with child
  • Requests for photos, especially "special" or private ones
  • Sexual language directed at children
  • Grooming behaviors (gifts, special attention, secrecy)

Violence Detection​

What we look for:

  • Threats of physical harm
  • Graphic violent content
  • Self-harm references
  • Dangerous activities or challenges

Risk factors that increase score:

  • Specific threats with details
  • Images of weapons or violence
  • Self-harm planning or methods
  • Encouragement of dangerous activities

Hate Speech Detection​

What we look for:

  • Discriminatory language based on identity
  • Hate symbols or references
  • Organized hate group activity
  • Targeting based on race, religion, gender, etc.

Risk factors that increase score:

  • Systematic targeting of individuals
  • Use of known hate symbols or terminology
  • Encouragement of discrimination or violence
  • Recruitment for hate groups

Context Factors​

Platform-Specific Considerations​

Gaming Platforms (Roblox, Minecraft):

  • Competitive language is common and scored lower
  • Focus on interactions outside of normal gameplay
  • Private messages weighted more heavily than public chat

Social Media (Instagram, TikTok):

  • Public posts get additional scrutiny
  • Comments on photos analyzed for appropriateness
  • Direct messages prioritized for safety

Messaging Apps (Discord, WhatsApp):

  • Private conversations carry higher weight
  • Group dynamics and peer pressure considered
  • Adult-child interactions flagged more aggressively

Age-Appropriate Adjustments​

Younger Children (Under 10):

  • Lower tolerance for any concerning content
  • Educational conversations about online behavior
  • Focus on stranger danger and sharing information

Tweens (10-13):

  • Balance between independence and protection
  • Attention to social dynamics and peer pressure
  • Cyberbullying prevention and response

Teens (13+):

  • Respect for developing autonomy
  • Focus on serious safety threats
  • Preparation for adult online responsibility

How Scores Are Calculated​

AI Analysis Process​

  1. Content Preprocessing

    • Text is analyzed for meaning and context
    • Images scanned for inappropriate content
    • User interaction patterns considered

  2. Multi-Factor Assessment

    • Base content analysis (language, imagery)
    • Context evaluation (platform, relationship)
    • Historical pattern recognition
    • Risk escalation factors

  3. Score Assignment

    • Weighted scoring based on severity
    • Platform-specific adjustments
    • Age-appropriate considerations
    • Final risk level determination

Learning and Improvement​

Continuous Enhancement:

  • AI learns from new online safety threats
  • Regular updates to detection algorithms
  • Parent feedback helps refine accuracy
  • Integration of latest research on child safety

Real Examples​

Low Risk Examples​

Score 2 - Gaming Chat:

"gg everyone, that was fun!"

  • Positive social interaction
  • Appropriate gaming language
  • No safety concerns

Score 3 - Social Media:

"Can't wait for the weekend! 😊"

  • Normal personal sharing
  • Age-appropriate content
  • Standard social media interaction

Medium Risk Examples​

Score 5 - Mild Conflict:

"You're so bad at this game, just quit"

  • Mild negative behavior
  • Could hurt feelings but not dangerous
  • Opportunity for teaching moment about kindness

Score 6 - Information Sharing:

"I go to Lincoln Middle School in the 7th grade"

  • Sharing personal information publicly
  • Not immediately dangerous but risky
  • Needs conversation about privacy

High Risk Examples​

Score 8 - Inappropriate Request:

"Send me a photo of yourself, don't tell anyone"

  • Clear predatory behavior
  • Request for secrecy is major red flag
  • Immediate parent intervention required

Score 10 - Direct Threat:

"I know where you live, I'm coming to hurt you"

  • Specific threat with apparent capability
  • Immediate safety concern
  • Law enforcement contact may be necessary

Taking Action Based on Scores​

Low Risk Response​

  • Monitor trends over time
  • Casual conversations about online experiences
  • Positive reinforcement for good digital citizenship

Medium Risk Response​

  • Review full context of the interaction
  • Talk with your child about what happened
  • Provide guidance on better responses
  • Monitor more closely for related incidents

High Risk Response​

  • Document evidence immediately
  • Ensure child's immediate safety
  • Contact appropriate authorities if needed
  • Restrict access to dangerous platforms/contacts
  • Seek professional help if trauma occurred

Frequently Asked Questions​

"Why did harmless content get a medium score?"​

Sometimes context matters more than the specific words. Our AI considers:

  • Who is sending the message (adult vs. peer)
  • Platform where it occurred (public vs. private)
  • Pattern of behavior (isolated vs. repeated)
  • Your child's age and vulnerability factors

"Can I adjust the sensitivity of scoring?"​

Yes! In your dashboard settings, you can:

  • Adjust notification thresholds
  • Focus on specific content types
  • Customize platform-specific settings
  • Set age-appropriate sensitivity levels

"What if I disagree with a score?"​

  • Use the feedback option on each incident
  • Your input helps improve the AI accuracy
  • Consider if additional context might affect the assessment
  • Contact support for scoring questions

Best Practices​

Daily Monitoring​

  • Focus on high-risk incidents first
  • Scan medium-risk content for patterns
  • Use low-risk data to understand normal activity

Weekly Reviews​

  • Look for trends in scoring over time
  • Identify platforms that generate more concerns
  • Adjust settings based on your family's needs

Teaching Moments​

  • Use incidents as conversation starters
  • Explain why certain content is concerning
  • Help your child develop good judgment
  • Celebrate positive online interactions

Next Steps​

Understanding risk scores empowers better protection! Use this knowledge to prioritize your attention and have more effective conversations with your child about online safety.