Verification Process

Verification is the cornerstone of trust in Thamizhi. Our multi-layer system combines AI efficiency with human judgment to ensure accuracy.

The Three-Layer System

Layer 1: AI Pre-screening

Automatic processing within minutes

The AI pipeline performs initial verification:

• Content analysis: Checks for factual consistency
• Duplicate detection: Identifies similar stories
• Source verification: Validates referenced information
• Bias detection: Flags potential one-sided reporting
• Quality scoring: Rates overall reliability

What AI checks:

• Grammar and clarity
• Logical consistency
• Factual claims against known databases
• Image authenticity (reverse search)
• Link validity

Limitations:

• Cannot verify eyewitness accounts
• May miss cultural context
• Struggles with satire or irony
• Needs human judgment for complex issues

Layer 2: Community Review

24-48 hours of peer evaluation

Community members review submissions:

• Accuracy voting: Upvote/downvote on factual claims
• Comment feedback: Questions and suggestions
• Source verification: Community validates references
• Context addition: Members add background information
• Bias identification: Community flags potential issues

Community review process:

1. Submission appears in review queue
2. Qualified reviewers examine content
3. They vote on accuracy (1-5 scale)
4. Comments provide detailed feedback
5. Author can respond and revise
6. Final community vote determines outcome

Reviewer qualifications:

• Account age > 30 days
• Reputation score > 70
• Active participation history
• No recent violations

Layer 3: Expert Validation

Specialized review for complex topics

Subject matter experts verify:

• Technical accuracy: For scientific or technical claims
• Legal compliance: For legal or regulatory content
• Medical accuracy: For health-related information
• Historical facts: For historical claims
• Cultural context: For cultural or social topics

Expert network:

• Verified professionals in various fields
• Academic researchers
• Industry specialists
• Community leaders with domain expertise

When expert validation is triggered:

• High-impact stories
• Technical or scientific claims
• Legal or medical content
• Controversial topics
• Community request

Verification Status

Every story shows its verification level:

Status	Meaning	Trust Level
🟡 Pending	Awaiting review	Low
🔵 AI Verified	Passed AI screening	Medium
🟢 Community Verified	Community approved	High
🟣 Expert Verified	Expert validated	Very High
🔴 Disputed	Accuracy questioned	Variable

The Verification Timeline

Submission → AI Screening (minutes) → Community Review (24-48h) → Publication
                                    ↓
                              Expert Validation (if needed)
                                    ↓
                              Final Verification

Factors Affecting Verification Speed

Faster verification:

• Clear, well-structured submissions
• Multiple credible sources
• Author with good reputation
• Straightforward factual content
• No controversial claims

Slower verification:

• Complex or technical content
• Controversial or sensitive topics
• Author with limited history
• Need for expert validation
• Community disagreement

After Verification

Verified Stories

• Appear in main feed with verification badge
• Higher visibility in search results
• Contribute to author’s reputation
• Eligible for platform promotion

Disputed Stories

• Flagged with dispute notice
• Author can appeal or revise
• Community discussion encouraged
• May be removed if unresolvable

Improving Verification Outcomes

For authors:

• Provide multiple sources
• Include context and background
• Respond to community feedback
• Build reputation over time
• Be transparent about limitations

For reviewers:

• Focus on facts, not opinions
• Provide constructive feedback
• Vote consistently
• Report genuine issues
• Help authors improve

Verification Metrics

Track your verification performance:

• Approval rate: Percentage of submissions verified
• Average review time: How quickly your content is reviewed
• Community feedback score: Average rating from reviewers
• Expert validation rate: How often expert review is needed

Common Verification Issues

Q: Why was my submission rejected? A: Check the feedback for specific reasons. Common issues include insufficient sources, factual errors, or bias.

Q: Can I appeal a verification decision? A: Yes, you can appeal through the dispute process. Provide additional evidence or context.

Q: How do I become a qualified reviewer? A: Build your reputation through accurate submissions and helpful community engagement.

Q: What if I disagree with community verification? A: You can flag the story for expert review or provide additional evidence in comments.

Next Steps

• Learn about Reputation System to understand trust scoring
• Read Community Guidelines for detailed rules
• Explore AI Pipeline for technical details