About the Project
The constant flood of sensationalism in the news wasn’t just frustrating — it was dangerous. I noticed how headlines were increasingly weaponized: not to inform, but to provoke, mislead, and manipulate. That was the starting point.
InfoShield was built as a direct response to that. A browser extension that doesn't just detect emotionally manipulative language — it flags it and backs it up with credibility scores so users can see the bias behind the words, not just the words themselves.
I built this using vanilla JavaScript and Chrome Extension APIs. I trained a simple keyword-based classifier backed by curated bias and sentiment datasets. The UI is minimal and purpose-driven — it's not meant to be flashy, just brutally clear.
What I Learned
Building this taught me how easily perception can be shaped with just a few emotionally charged words. On the technical side, I dove deep into browser extension architecture, NLP basics, and DOM parsing.
Challenges
Not all patterns of bias are easy to catch. Creating a lightweight but meaningful detector meant finding a balance between simplicity and impact. Handling dynamic web content (like Al Jazeera’s) also threw rendering bugs I had to resolve with smarter DOM scanning and regex handling.
In short: I wanted to give users a tool to reclaim control over how they interpret information — and keep the noise out.
Log in or sign up for Devpost to join the conversation.