AI Factchecker

Inspiration

Fake news and misinformation is a current problem in the world and is very dangerous

What it does

AI FactChecker allows users to input statements which the AI will check using APIs like google and wikipedia to check and return if the statement is true or false

How we built it

Frontend: Next.js (React), TailwindCSS Backend: FastAPI, Hugging Face Transformers, Wikipedia API, Google APIs ML Model: microsoft/deberta-large-mnli (Zero-Shot Classification) APIs Used: Google Fact Check API (for claim verification) Google Custom Search API (for retrieving news articles) Wikipedia API (for knowledge-based verification)

Challenges we ran into

Trying to get the AI to be more accurate

Accomplishments that we're proud of

AI-Powered Claim Verification - Uses zero-shot classification models Google Fact Check API Integration - Cross-checks claims with trusted fact-checking sources Wikipedia Verification - Fetches summaries for evidence-backed analysis Google Search Backup - Retrieves real-time news articles if fact-checking is unavailable FastAPI Backend + Next.js Frontend - High-performance, scalable infrastructure

What we learned

Creating an AI is difficult

Built With

• api
• apis
• articles)
• check
• claim
• classification)
• custom
• fact
• for
• frontend:-next.js-(react)
• google
• hugging-face-transformers
• knowledge-based
• microsoft/deberta-large-mnli
• ml
• model:
• news
• retrieving
• search
• tailwindcss-backend:-fastapi
• used:
• verification)
• wikipedia
• wikipedia-api
• zero-shot