Bias Buster-Ethical Ai Usage

Landing page
Biased Statement
Neutral Statement

Inspiration

With the rapid adoption of generative AI, we noticed a rising concern: AI responses can often carry unintentional bias or reflect skewed data. As developers and learners, we wanted to build a tool that promotes ethical AI usage by helping users detect and mitigate bias in AI-generated content.

What it does

Bias Buster allows users to paste any AI-generated response and receive an instant bias analysis powered by Google Gemini. The tool returns:

A verdict: Biased or Neutral
An explanation of the bias (if any)
A rewritten, more neutral version of the input

How we built it

We built the project using:

Streamlit for the web interface
Google Generative AI (Gemini API) for the bias analysis
Lottie animations for interactive visuals
TSParticles for a responsive, animated starry background

The app was styled using custom HTML/CSS injected into Streamlit, giving it a visually engaging, dark-themed interface.

Challenges we ran into

Integrating the Gemini API securely without exposing keys publicly
Streamlit's limitations with client-side JavaScript made animations tricky
Handling unexpected responses or timeouts from the AI model
Managing deployment and environment configuration via GitHub and Streamlit Cloud

Accomplishments that we're proud of

Successfully built and deployed a working MVP with a clean UI and seamless user flow
Learned how to integrate real-time LLM responses into a web app
Created a tool that addresses a real and growing problem in AI adoption