BiasBeGone

Inspiration

The spread of biased/fake news and misinformation has skyrocketed throughout the years, making it hard to distinguish from unbiased/real information.

The issue often disproportionately affects older adults compared to other age groups, usually taking place on social media.

Biased/fake news is often strategically designed to influence those who are less media-literate and/or have less experience to social media and biased content, thus creating an element of strategic bias.

Our Solution?

We created an easy-to-use chrome extension that can read the page you are on and determine if the content is biased or fake content.

How it works:

• First it scrapes the html of your current page (say a news article)
• Then, we query the title with similar articles on Google for a comparison
• We pass both the text on your current page and related articles to an LLM for comparison
• The LLM is instructed with a system prompt to tell if you if the current article/text you are reading might be biased/fake.
  

How we built it

Chrome Extension Frontend: JavaScript. Text Processing: Langchain + Gemini 2.5 Flash. Web Scraping: BeautifulSoup4, Custom Google Search API. Backend: FastAPI.

Challenges we ran into

• Using APIs and dealing with back-end issues!
• Figuring out how to connect all of our pieces of code together!

Accomplishments that we're proud of and what we leanred!

• Learned a lot more about web-development.
• Used FastAPI to create an endpoint.
• Got more familiar with Git and Github.
• Managed to submit something, despite changing ideas last-minute!

What's next for BiasBeGone

• A more carefully designed fake/bias detection system.
• Re-designing the UI elements.

Built With

• fastapi
• gemini
• googlesearch
• javascript
• langchain
• python