-
-
AI-powered platform for detecting sensitive data leaks.
-
End-to-end pipeline for identifying sensitive information
-
Unified workflow for secure AI usage and data protection
-
Detects privacy risks and sensitive information in prompts.
-
Scans PDFs to identify sensitive information.
-
Highlights detected sensitive information and security risks.
-
Centralized dashboard for monitoring security events.
-
Real-time protection against AI-related data leaks.
-
Quick access to project resources.
Inspiration
As organizations increasingly use AI tools and AI agents, employees may unintentionally share confidential information such as credentials, personal data, source code, or business documents. Shadow-AI was created to address this problem by helping organizations detect and prevent sensitive data exposure during AI usage.
What it does
Shadow-AI is an AI security platform that detects unauthorized AI usage and sensitive information leaks. It analyzes prompts and documents, identifies privacy risks, and provides security insights through a centralized dashboard. The platform also includes a browser extension for real-time protection.
How we built it
We built Shadow-AI using AI/ML and web technologies to create an end-to-end security workflow. The system analyzes prompts and uploaded documents, detects sensitive information, evaluates privacy risks, and displays the results through an interactive web interface and monitoring dashboard. We also developed a browser-based security layer for real-time detection.
Challenges we ran into
One of the main challenges was accurately identifying sensitive information while minimizing false positives. We also had to design the detection pipeline to work across different types of prompts and documents while keeping the system practical for real-time usage.
Accomplishments that we're proud of
We are proud of building a complete AI security workflow that combines sensitive-data detection, prompt analysis, PDF scanning, browser protection, and centralized monitoring into one platform. We also successfully developed a functional product with multiple integrated components.
What we learned
Through Shadow-AI, we gained practical experience in AI/ML integration, prompt analysis, data privacy, full-stack development, API integration, and building security-focused AI applications. We also learned how to collaborate effectively as a team while developing a complete product.
What's next for Shadow-AI
We plan to improve detection accuracy, expand support for different AI platforms and data types, strengthen real-time monitoring, and introduce more advanced AI-powered privacy and security capabilities.
Log in or sign up for Devpost to join the conversation.