Bringing Our Hate-Speech Detection System to Life —Usability Study and System Demonstration Recap
On October 30th, the SOCYTI team hosted our second major usability study — a milestone moment in validating the technology behind our hate-speech detection and prevention assistance system. This time, we were joined by five incredible external partners representing anti-hate activism NGOs and LGBT advocacy groups. Their expertise and feedback are vital to shaping a tool that truly works for those on the frontlines of combating online hate.
What Was the Goal?
The purpose of this study was simple yet ambitious: to showcase a fully functional build of the system that reflects the final product vision. Compared to our first prototype last August, this latest version featured:
A revamped user interface for smoother navigation.
Refined model retraining options, giving users more control.
Significantly improved AI performance, thanks to months of optimization.
How Did It Work?
We kicked off with a guided demo, walking participants through the system’s core features and providing a detailed user manual for reference. After a quick coffee break, it was time for hands-on exploration.
SOCYTI tool interface
To make the experience realistic, we designed a roleplay scenario: participants stepped into the shoes of data analysts tasked with identifying and mitigating hate speech using SOCYTI. This exercise mirrored real-world workflows and tested how intuitive the tool feels in practice.
Collecting Feedback
After the hands-on session, we gathered insights in two stages:
Questionnaire: Focused on task completion, clarity of data visualizations, and overall usability.
Discussion: A long-form conversation where participants shared what worked, what didn’t, and what they’d love to see next.
The feedback was sharp, constructive, and energizing. Our partners didn’t hold back—and that’s exactly what we needed.
What’s Next?
We’re thrilled with the progress and even more excited about what’s coming. The final usability study is scheduled for March 15, 2026, where we’ll present the fully polished version of the system; complete with all the improvements inspired by feedback from this and previous sessions.
A huge thank you to our stakeholder attendees and SOCYTI team members for their time, insights, and commitment to making SOCYTI a powerful ally in the fight against online hate. Together, we’re building something that matters.