Deepshield

• 

  Detect Media → “DeepShield Detection — Upload media, uncover deepfakes and AI‑generated content instantly.”

• 

  Cases → “DeepShield Cases — Authenticity logs and escalated deepfake alerts tracked transparently.”

• 

  File Report → “DeepShield Reporting — Secure, masked submissions to cyber crime authorities.”

• 

  Evidence Chain → “DeepShield Evidence — Tamper‑proof SHA‑256 chain ensuring legal admissibility.”

• 

  Dashboard → “DeepShield Dashboard — Real‑time detection metrics and trust at a glance.”

✨ Inspiration

The rise of deepfakes and AI-generated media is eroding public trust in digital content. Governments, journalists, and citizens often lack reliable tools to verify authenticity. This inspired me to build DeepShield — a platform that combines detection, case management, reporting, and tamper-proof evidence chains to restore digital trust.

🛠️ How I Built It

• Backend: Node.js with Express for handling uploads and API calls.
  
• Detection Engine: Initially integrated Sightengine APIs for quick prototyping.
  
• Dashboard & Case Management: JSON-based storage (cases.json) with live metrics and case tracking.
  
• Reporting Module: Secure submission to cyber crime authorities with masked identity.
  
• Evidence Chain: Implemented SHA-256 hash chaining to ensure tamper-proof verification:
  

Any alteration → hash mismatch → chain invalid.

📚 What I Learned

• How to integrate external APIs (Sightengine, and exploring Gemini).
  
• Importance of modular design: detection engine can be swapped without breaking the workflow.
  
• Basics of cryptographic hashing for evidence verification.
  
• How to frame technical work into a social impact narrative for hackathons.

⚔️ Challenges

• API Migration: Sightengine was easy to plug in, but moving to Gemini requires new endpoints and response parsing.
  
• Contributor Limits: GitHub repo permissions slowed deployment testing.
  
• Risk Management: Needed to avoid breaking the working prototype, so I experimented by copying the project folder before changes.
  
• Transparency vs Complexity: Balancing beginner-friendly design with advanced detection logic.

🌍 Impact

DeepShield is more than a prototype — it’s a step toward AI for Social Good. By making detection transparent and evidence legally admissible, it empowers governments, institutions, and citizens to fight misinformation and protect democracy.

Built With

• css3
• html5
• javascript
• sightengineapi