I use machine learning to reduce polarization in political discussions.
This bot was trained on a corpus of over 2 million tweets to recognize extremely politically polarized text. It does not discriminate between polarization regarding Democrats or Republicans, and is very non-biased in its moderation capabilities. This recognition module is actually the subject of a paper in-development, and it is the most advanced and accurate detection module to date. If it detects a user is getting very polarized in their speech, it will send them a warning for their first 3 very polarized statements, than mute them for increasing periods of time. This bot is intended to be used to foster healthy political discussion on a server, without removing the option for political discussion altogether.
