Sentinel AI is available as a Website, Chrome extension, and WhatsApp bot
AI Agent Debate – Input a statement, and AI agents will research, debate, and refine arguments based on user input, providing context to users about gray area topics because misinformation isn’t about right or wrong anymore, it can be about bias as well! (Never been done before)
YouTube Fact-Checking – Input a YouTube link, and our AI Agent (Llama, Google Fact Checker API, SerperAI, Web Scraping) conducts a double-layer fact check, providing a summary, context, and sources.
Text/Audio Fact-Checking – Works like the above but can also records Zoom calls for real-time verification.
(Users can also download reports from all checks)
Deepfake Detection:
Image: Triple check using CNN-based autoencoders, HuggingFace models, and VLMs
Video: VLM check and context check (e.g., has this person ever said this? Would they ever say this?).
Audio: SVM model
Scam Detector: – Detects scams in messages and emails using ML models (RoBERTa, XGBoost).
This project is highly technical, and innovative. It also has a significant impact on society, as it is highly accessible and covers a lot of the key problems facing Singaporeans and the world today. Also SentinelAI primarily uses FREE (open source) AI models - which means that our product is FREE and ACCESSIBLE by ALL!
Team: Pratham, Aditya, Alok, Akshar (Team 50), added a backup drive video link for demo!
Built With
- swag
Log in or sign up for Devpost to join the conversation.