Hackers Against Harassment

Inspiration

Drawing inspiration from the collaboration between MLH and #HackHarassment, we decided to create a social media bot that could differentiate offensive comments based on the measurement of key words said by a user. We wanted to promote integration with our bot so that it could be used with different platforms. Additionally, a supportive community is necessary to help those who have been affected by online harassment. Therefore, we sought this as an opportunity to remove negativity from online spaces as well as providing help for the victim.

What it does

In the context of Reddit, HAH can look at a subreddit and scan the top comments for harassment. It takes account for the number of times a key word is used by each user. A list of top users of inflammatory comments is marked on a list that is sent to the moderator of the subreddit. If a user has submitted too many mean comments, the moderator of the subreddit can determine whether the user can be shadow banned or not.

The bot does not ban the users based solely on key words, but assists moderators in determining recurring offenders. Human moderators are preferred to make the final decision, such as deleting the comment or banning a person, based on the context of the comment. Bots cannot clearly notice the difference between sarcasm and intentional abuse.