Resources in this category center the voices and experiences of T&S practitioners. This section includes videos of conference presentations, academic research about content moderators, toolkits and reports, blog posts written by T&S professionals, and podcasts featuring interviews with T&S leaders.
A voluntary and open source image classification system adopted by members of the Tech Coalition that is used by many electronic service providers to categorize images and videos that depict apparent child sexual abuse and exploitation.
This template has been created to help you get started on your company’s first CSEA transparency report. It aligns with TRUST: Voluntary Framework for Industry Transparency. The template provides suggestions and guidance for the three major sections of a report: Policies and Practices, Processes and Systems, and Outcomes. For additional information on transparency reporting best practices, please see the Tech Coalition’s Trust Framework - Transparency Reporting Implementation Guide.
The Tech Coalition’s Trust: Voluntary Framework for Industry Transparency (the Framework) provides principles- based guidance to tech companies seeking to build trust around their efforts to address online child sexual exploitation and abuse (CSEA) risks on their services. This guide serves to aid industry in implementing the Framework and moving toward better alignment in transparency reporting.
Trust: Voluntary Framework for Industry Transparency (the Framework) was developed by the Tech Coalition to provide principles-based guidance to tech companies seeking to build trust around their efforts to address online child sexual exploitation and abuse (CSEA) risks on their services. The Framework outlines principles that provide a general basis for considering how to approach transparency reporting and recommended report categories.
A collection of blog posts from the engineers at ActiveFence
TL;DR: “Precision” is a measure of the accuracy of a detection system. Improving precision is to reduce your false positive rate. “Recall” is a measure of the coverage of a detection system. Improving recall is to reduce your false negative rate.
Prior research on content moderators has failed to explore moderators' initial reactions to content from the start of employment through tenure as the subjection to material and habituation increases. This qualitative study takes an in-depth look at moderators' experiences from recruiting, through training, and production to better understand the content moderators' startle response and factors that enable startle habituation.