Analyze Webpages for Social Bias with Fair-ly
Fair-ly is a Chrome extension designed to analyze webpages for social bias, utilizing advanced AI models based on the latest research in bias detection. It offers features such as binary text classification to determine if content is biased or unbiased, as well as multi-label aspect classification to identify specific types of bias, including gender and racial bias. Additionally, it provides multi-label token classification, which labels words within sentences to highlight generalization, unfairness, and stereotypes.
This extension is part of the open-source Fair-ly Project and aims to make bias detection tools more accessible for researchers and developers. Users can explore the underlying research papers associated with the models through the project's documentation. Furthermore, a Python package is available for those who wish to implement model inference in their coding projects, enhancing the utility of bias analysis in various applications.
