Skip to content

🚀 ChainShield

🗂️ Project Overview

  • Student ID: 20240562
  • Name: Joonseok Lee
  • Project Title: ChainShield
  • Summary (3–4 sentences):
    ChainShield is an AI-powered platform that detects and analyzes cybercrimes—such as phishing, scams, and malicious links—and records verified incidents on the blockchain. Users can upload suspicious messages or URLs, which are then analyzed using large language models and threat classifiers. Verified cases are stored on-chain to create a transparent, immutable, and censorship-resistant public cybercrime registry. This combination of AI and blockchain helps individuals and organizations prevent repeated attacks and enables researchers and journalists to track emerging threats.

1. 🧩 Problem: What Problem Are You Solving?

  • Problem Statement: Cybercrimes like phishing, fake websites, and deepfake-based scams are increasing rapidly, especially as generative AI makes these attacks more convincing and scalable. Victims often don’t report these incidents or don’t have a secure, trusted place to do so. Even when reports are made, they are typically siloed within private platforms or law enforcement and are not publicly verifiable.
  • Limitations of Existing Solutions:
  • Most reports are deleted or inaccessible to the public.
  • There is no decentralized, tamper-proof record of incidents.
  • AI-driven cyberattacks outpace traditional security filters.
  • Why This Problem Matters: Preventing cybercrime and raising public awareness requires transparency. A shared, trustworthy record of attack patterns empowers journalists, researchers, developers, and everyday users to defend themselves better.

2. 💡 Solution: Your Proposed Approach

  • Proposed Solution: ChainShield allows users to submit suspicious content (e.g., a phishing message, scam link, deepfake video). AI models analyze the content, classify the type of cybercrime, and extract relevant metadata. If confirmed, a public, hashed summary of the attack is published to a blockchain network like Base or Solana. This registry can be queried by anyone, or accessed via a public API.
  • Combining AI
  • LLMs (like GPT-4 or Claude) for semantic understanding of suspicious text
  • Specialized classifiers for phishing, scam, and deepfake detection
  • Similarity search with vector embeddings for prior case matching
  • Combining Blockchain
  • Immutable publishing of attack summaries and evidence hashes
  • On-chain registry serves as a community-owned, censorship-resistant threat intelligence hub

3. 🔗 Why Blockchain (and Token)?

  • Why Blockchain:
  • Ensures transparency and immutability of cybercrime incident records
  • Allows the public to verify reports independently
  • Prevents platform censorship or data loss over time
  • Enables reputation systems for reporting agents and high-trust reporters
  • Token Design: A Reputation token is designed for trusted reporters or incentive mechanisms in crowd-sourced moderation and case review.

4. 🛠️ MVP or Prototype

  • Current status:
    ☑ Idea only ☐ Prototype ☐ Working MVP

~~5. 📬 Submission to Hackathons or Grant Programs~~ (Exempted)

6. 🤔 Reflection & Future Work

  • Biggest Challenges:
  • Designing robust AI pipelines for multiple forms of cybercrime
  • Keeping blockchain cost low for frequent submissions
  • Avoiding abuse or false reporting in a decentralized system
  • Real-world cybercrime analysis requires a careful balance between automated classification and false-positive control.
  • Future Improvements:
  • Add vector-based search for similar past attacks
  • Expand to support deepfake video and image detection
  • Create browser plugin for real-time threat warning

7. 📚 References