Using the AI Verification Service
Complete workflow for submitting and reviewing verifications
Overview
The AI Verification Service integrates seamlessly with the CopyrightChains Dashboard, allowing administrators to verify copyright registrations with a single click. This guide walks through the complete verification workflow from start to finish.
Dashboard Workflow
The primary method for using the verification service is through the CopyrightChains Dashboard. This provides the most streamlined experience with automatic session handling and report display.
Step 1: Access Dashboard
Navigate to the CopyrightChains Dashboard and connect your wallet with SempreID authentication. Your administrator status will be verified automatically.
- URL: https://dashboard.copyrightchains.com
- Authentication: MetaMask wallet with valid SempreID
- Access Level: Administrator privileges required
Step 2: Review Pending Registration
Navigate to the admin panel and select a pending registration to review. The registration details page displays all submitted content including audio files, lyrics, metadata, and contributor information.
Available Information:
- Musical work title and type (composition or recording)
- Submitted audio files and format details
- Lyrics and score sheets if provided
- Contributor wallet addresses and split percentages
- Submission timestamp and applicant information
Step 3: Trigger AI Verification
Click the "Run AI Verification" button on the registration detail page. The Dashboard will send a verification request to the AI service with the registration ID and all associated content.
- Button located in admin actions section
- Single click starts verification process
- No manual file upload required
- Automatic content retrieval from database
Step 4: Monitor Progress
A progress indicator appears showing the verification status. Processing typically takes 12-60 seconds depending on content complexity and provider availability.
Progress States:
- Queued: Verification added to processing queue
- Processing: AI providers analyzing content
- Completed: Report ready for review
- Failed: Error occurred, check logs or retry
Step 5: Review Report
Once completed, the verification report appears on the registration page. Review the overall risk score, individual confidence scores, and detailed findings from each analysis type.
- Executive summary with recommendation
- Individual scores for each verification type
- Detailed findings and evidence
- Provider-specific analysis results
Step 6: Make Decision
Based on the verification report, approve or reject the registration. The report provides a recommendation, but final decision rests with the administrator.
Decision Guidelines
- Low risk (0-30): Generally safe to approve
- Medium risk (31-70): Manual review and investigation required
- High risk (71-100): Recommended rejection or extensive investigation
API Integration
For advanced users and custom integrations, the AI Verification Service provides a REST API. This allows programmatic access to verification functionality outside the Dashboard.
API Endpoint: Submit Verification
POST /api/v1/verify
Authentication: Required (SempreID session cookie)
{
"registrationId": "uuid-of-registration",
"type": "musical_work",
"content": {
"title": "Song Title",
"files": {
"audio": "base64-encoded-audio-data",
"lyrics": "base64-encoded-lyrics",
"score": "base64-encoded-score"
},
"metadata": {
"duration": 180,
"format": "mp3",
"bitrate": 320
}
}
}
Response:
{
"success": true,
"verificationId": "generated-uuid",
"status": "queued",
"estimatedTime": 60
}
API Endpoint: Check Status
GET /api/v1/status/:verificationId
Authentication: Required
{
"verificationId": "uuid",
"status": "processing",
"progress": 75,
"currentStep": "Analyzing audio fingerprints",
"updated_at": "2025-10-16T12:00:00Z"
}
API Endpoint: Retrieve Report
GET /api/v1/report/:verificationId
Authentication: Required
{
"verificationId": "uuid",
"registrationId": "uuid",
"type": "musical_work",
"status": "completed",
"report": {
"overallRisk": 15,
"infringementRisk": 10,
"launderingRisk": 5,
"authenticityScore": 95,
"duplicateProbability": 0,
"recommendation": "approve",
"findings": { ... },
"providerResults": [ ... ]
}
}
API Authentication:
- Session-based authentication using secure cookies
- SempreID validation on every request
- Administrator status verified automatically
- No manual token management required
Batch Verification
When reviewing multiple registrations, administrators can queue several verifications in sequence. Each verification is processed independently with its own report.
Batch Processing Workflow
- Navigate to pending registrations list
- Select multiple registrations for verification
- Click "Batch Verify" button
- Verifications added to queue sequentially
- Monitor queue status in admin panel
- Review reports as they complete
- Make bulk approval/rejection decisions
Queue Performance:
- Maximum 5 concurrent verifications
- Additional requests queued automatically
- Average processing time: 12-60 seconds per verification
- Priority processing not currently supported
Common Workflows
Workflow 1: Standard Registration Review
Most common use case for single registration verification:
- Receive notification of new registration
- Access Dashboard and navigate to pending items
- Review submitted content and metadata
- Click "Run AI Verification"
- Wait 12-60 seconds for processing
- Review generated report
- Make approval/rejection decision
- If approved, proceed with blockchain registration
Workflow 2: Suspicious Content Investigation
For registrations flagged as potentially problematic:
- Run initial AI verification
- Review report for specific concerns
- If medium/high risk, examine detailed findings
- Check provider-specific analysis
- Cross-reference with external databases if needed
- Request additional evidence from applicant
- Re-run verification if new content provided
- Make informed final decision
Workflow 3: Bulk Review Session
Processing multiple registrations efficiently:
- Filter pending registrations by date or type
- Submit batch verification for all selected items
- Monitor queue progress in real-time
- Review reports as they complete
- Flag high-risk items for deeper investigation
- Batch approve low-risk registrations
- Schedule follow-up for medium-risk items
Verification Retry
If a verification fails or produces incomplete results, you can retry the verification. Common reasons for retry include provider timeouts, network errors, or incomplete content submission.
When to Retry
- Timeout Errors: Provider took too long to respond
- Partial Results: Some providers failed but others succeeded
- Network Issues: Temporary connectivity problems
- Provider Outages: Specific AI provider temporarily unavailable
Retry Process
- Review error message in failed verification
- Check service health status
- Wait 2-5 minutes if provider outage suspected
- Click "Retry Verification" button
- Monitor new verification progress
- Compare results with previous attempt if partially successful
Best Practices
Recommended Practices
- Always run verification before final approval decisions
- Review full report, not just overall risk score
- Pay special attention to medium-risk results
- Document reasoning for decisions contradicting AI recommendations
- Verify all files are submitted before triggering verification
- Check service health before large batch operations
- Monitor queue status during peak usage periods
Things to Avoid
- Don't approve high-risk registrations without investigation
- Don't skip verification for "trusted" submitters
- Don't rely solely on AI scores for final decisions
- Don't submit verifications for incomplete registrations