Project for a community-based and automated tool for the detection and demonetization of hate speech

AMALIA: Community-based, automated tool project for the detection and demonetization of hate speech. We are part of a community of associations, digital activists, content creators, and journalists concerned about how online hate speech takes advantage of the attention economy.

Our goal is to deplatform hateful content from platforms and demonetize it to discourage its creators from continuing to produce it. To achieve this, we follow the watchdog journalism and "follow the money" approach implemented by organizations like Sleeping Giants , Check My Ads or ÉsPaísValencià.

One of the first needs expressed by our community was to automate the detection of hate speech across different social media platforms. This task was highly detrimental to the emotional well-being of those who volunteered to perform it, as they had to consume countless hours of content rife with hate speech. That's why we decided to create AMALIA: an automated tool for detecting and demonetizing hate speech.

We will discuss the decisions and stages we have gone through to define the dimensions and categories that allow us to identify hate speech, creating an inclusive and balanced sample, with the communities targeted by hate speech at the helm. We will explain the current state of the tool, as well as the difficulties we face in achieving effective reliability.