Abstract

The Um Actually! project aims to address the growing spread of biased information across social media and other online platforms. As more people rely on the internet for news and information, it has become increasingly difficult to determine what sources are trustworthy. The Um Actually! extension was developed to be compatible with all major desktop web browsers and enabled users to use AI and identify biases within texts anywhere on the Internet as well as the user's own writing. The extension uses Google's Gemini, Firebase and WXT to implement, and evaluate all text. Users were also encouraged to engage with the project through game-like features to encourage continued use of the extension and to see how their ability to detect bias changed over time. Along with identifying biases, our project emphasizes user education by offering feedback and clear explanations to help users understand and develop their bias recognition skills.