logo

InfoShield: Your AI-Powered Digital Lie Detector

Inspiration

In an era where misinformation spreads faster than the truth, we wanted to create a data-driven tool that empowers users to instantly verify online content—without the need for deep research. From manipulated headlines to deepfake images, misinformation has real-world consequences, influencing elections, public health, and everyday decision-making. Inspired by the need for trustworthy digital spaces, we built InfoShield: a browser extension that highlights potential misinformation and provides reliable sources at a glance.

What It Does

InfoShield acts as a real-time misinformation filter, analyzing the credibility of online content using AI-powered text processing. Key features include:

Website-wide analysis – Instantly scans an entire webpage and flags major issues and potentially false claims with a credibility score and reasoning.
Text highlighting verification – Users can highlight specific text to get a credibility score, true/false verdict, and a one-sentence reasoning with references to verified, credible sources.
Source referencing – Provides links to accurate sources for easy verification.
Seamless browser integration – Works in the background while users browse the web.

How We Built It

We combined React and Node.js for a dynamic frontend and backend, ensuring a smooth user experience. For misinformation detection, we leveraged:

OpenAI's NLP models to analyze and classify content.
FastAPI & Python scripts for processing text-based misinformation.
Custom credibility scoring based on cross-referencing reputable sources.
Browser extension integration for real-time analysis and content underlining.

Challenges We Ran Into

Balancing speed & accuracy – Processing large amounts of text efficiently while maintaining accuracy was a key challenge.
Training AI to detect misinformation – Since misinformation is nuanced, building a reliable detection model required careful tuning and dataset selection.
Frontend/backend integration – Ensuring smooth communication between the browser extension, backend APIs, and AI models was trickier than expected.

Accomplishments That We're Proud Of

Functional prototype – We built a fully working browser extension that successfully detects and flags misinformation.
AI-powered credibility scoring – Our integration with OpenAI provides concise explanations and alternative sources.
Real-time misinformation verdict – Users get instant feedback on online claims without leaving the page.

What We Learned

AI isn't perfect—but it’s powerful – While no misinformation model is 100% accurate, we learned how to fine-tune AI to provide meaningful and actionable insights.
Seamless UX is key – Fact-checking needs to be fast and intuitive; users won’t engage with a tool that disrupts their browsing experience.
The misinformation problem is vast – Tackling text-based misinformation is just the beginning—images, videos, and deepfakes require even more advanced solutions.

What's Next for InfoShield?

Expanding beyond text – Integrating image and video analysis to detect deepfakes and manipulated media.
User feedback loop – Allowing users to report incorrect classifications and help refine our AI.
B2B partnerships – Exploring integrations with news platforms, educational institutions, and fact-checking organizations to expand InfoShield’s impact.