Inspiration
The rapid rise of phishing attacks, scam messages, and AI-generated fraudulent content inspired us to build AI SHIELD. With increasing digital adoption in Bharat, many users fall victim to financial fraud, fake job offers
What it does
AI SHIELD is an AI-powered web application that analyzes suspicious messages and detects potential fraud or phishing attempts. Users can paste any message into the system, and the platform evaluates it using a large language model to:
- Identify scam indicators such as urgency, fake links, and social engineering tactics
- Generate a risk score (0–100%)
- Classify the threat level as Low, Medium, or High
- Provide a clear explanation of why the message is suspicious
- Maintain a history of recent checks
The goal is to help users quickly assess whether a message is safe or potentially harmful.
How we built it
AI SHIELD was built as a full-stack web application.
Frontend:
- Developed using HTML, CSS, and JavaScript
- Designed a cyber-themed, glassmorphism UI for better visual impact
- Implemented dynamic risk meter, badge system, loader animation, and history panel
Backend:
- Built using Node.js and Express.js
- Created REST API endpoint to receive user messages
- Integrated Google Gemini API for AI-based content analysis
AI Logic:
- Structured prompt engineering to guide the AI model to detect fraud patterns
- Processed AI responses and mapped them to visual risk indicators
Challenges we ran into
- Handling API version and model compatibility issues during Gemini integration
- Designing a meaningful risk scoring approach from natural language responses
- Managing real-time UI updates for a smooth user experience
- Ensuring clear, non-technical explanations for end users
- Balancing performance with accurate AI analysis
Accomplishments that we're proud of
- Successfully integrated AI into a real-world fraud detection use case
- Built a working end-to-end MVP with both backend and frontend
- Designed a visually impactful cybersecurity-themed interface
- Implemented dynamic risk visualization instead of plain text output
- Created a scalable structure that can evolve into a larger security platform
What we learned
- Practical integration of generative AI APIs into production-like systems
- Importance of prompt engineering in improving AI reliability
- Real-world challenges in cybersecurity-focused product design
- How UI/UX plays a critical role in user trust for security tools
- Effective collaboration through clear role division and structured development
What's next for AI SHIELD – Bharat’s Intelligent Fraud
- Real-time phishing URL validation using domain reputation APIs
- Automated suspicious keyword highlighting
- PDF-based downloadable security reports
- Cloud deployment for public access
- Browser extension for real-time message scanning
- Expansion into voice and deepfake content detection
- Multilingual support to serve users across Bharat
AI SHIELD aims to evolve into a comprehensive AI-driven fraud prevention ecosystem for individuals and small businesses.
Log in or sign up for Devpost to join the conversation.