Case study

Pride: The Power of Moderation in Communicating Sensitive Topics

September 2, 2024

The topic of “Pride,” which celebrates diversity and the LGBTQIA+ community, still receives a huge amount of hate, both in real life and on the internet. At elv.ai, we analyzed nearly 15,800 comments on the subject throughout July 2024 to understand public discourse and determine the most effective ways to engage with it. Let’s take a look at what Pride online discussions would look like if they weren’t moderated. 🏳‍🌈


Table of contents:

  1. Discussions without vs. with Moderation
  2. Does Online Hate Speech Have Real-World Impact?
  3. How to Deal with Online Hate Speech
  4. Differences between Platforms and Countries
  5. Sources Overview

Without moderation, anger would outweigh kindness in discussions 📊

The graph above shows a comparison of all the discussions examined unmoderated vs. after they were moderated by elv.ai. As can be seen, after moderation, the discussions contained 62.12% less anger. At the same time, they contained 26.81% more happiness and 35.76% more kindness.

Comments would also look like this without moderation.

Surprisingly, even before moderation, positive sentiment outweighed negative sentiment in the comment sections. However, moderation has managed to boost the lead of positive sentiment even further, specifically by increasing the difference from the original 0.005 to 0.015, a difference of 181.31%.

Does Online Hate Speech Have Real-World Impact? 🤔

Online hate is reflected in increased crime rates in the offline world. According to data from a study by the University of Warwick, anti-refugee sentiment on the internet had a significant impact on the rate of hate crimes against refugees in Germany. For every 4 anti-refugee posts on the internet, there was one corresponding “offline” hate crime. The correlation was most significant in places with higher Facebook usage, highlighting the impacts that social media have on our perception of the real world.

How to Deal with Online Hate Speech

People tend to post more hate speech if they can see some already posted. After 4 months of moderating with elv.ai, our clients have seen a 35% increase in the number of polite and positive comments in their discussions, while the number of toxic comments decreased by 46%. Your audience likes to engage in a safe environment where backlash is not tolerated.

“We respond to emotionally charged comments/opinions on issues and misleading and false questions in a polite, calm, factual manner, asking the commenters questions back, for example where they are getting their information from, trying to contradict their opinion or view.”

Prague Pride

Prague Pride offers the option to report bias attacks on their website. For more information on what a bias attack is and a form to report it, please visit the website here.

Facebook as the Platform with the Highest Hate Rate 🤬

Unsurprisingly, the infamous Facebook discussions have taken the top spot in the hate rate here as well. The hide rate for toxic content was as high as 34%. Compared to web discussions, this is only a few percent ahead, but it outperformed Instagram discussions by as much as 13%.

It was surprising that, despite the lower number of comments in the Czech language, we observed in them a toxicity rate 10% higher than in the Slovak ones.

Sources Overview 🌈

In total, we analyzed approximately 370 posts and 15,765 comments posted on the topic of Pride on accounts moderated by elv.ai since July. And to be as diverse as the rainbow, we took several platforms into consideration:

  • Facebook: 10,920 comments
  • Instagram: 4,398 comments
  • Web discussions: 1,504 comments

11,623 comments (73.7%) were in Slovak, 4,142 comments (26.3%) in Czech.
Of all these comments, we hid a whopping 34% of comments due to toxic content. The most commented of the examined posts amounted to almost 1,900 comments, but 32.46% of them had to be hidden.
One post, however, took our breath away: it reached an unprecedented toxicity rate of 66.3%. And that’s with a total of only 546 comments. 

So What Can We Do?

Hate speech of any kind is simply not acceptable in modern society. It’s important to listen to public discourse online and offline, as we’ve seen hate speech manifest in tragedies many times already, and listening is the first step to their prevention.

If you see or experience hate speech, don’t hesitate to report it. Repercussions are crucial at such moments and if you’re not able to get rid of the content, you should inform the platform as soon as possible.

Let’s listen to and moderate discussions and interact with our audience. Let’s make it clear that the space below our posts is not a space for bullying, but for education. In doing so, we protect the affected groups, but also the image of our brand and society as a whole.


Bring humanity and decency back to your discussions! 💌

With elv.ai’s content moderation tools, you don’t need to worry about hate speech flooding your discussions, even in a crisis or at a sensitive time. Our combination of AI and human moderators ensures that all hate comments, even the creative ones, are removed to make space for a safe and polite environment. Request a demo today and try it out for free.

Other case studies

October 21, 2024
CNN Prima Saves 8,500€ Monthly on Content Moderation With...
October 18, 2024
How Forbes Increased Engagement and Protected Brand Image on...
May 19, 2024
After the shooting in Prague, social networks were flooded...

Save 30% of Your Social Media Manager's Time

No credit card required, cancel anytime.

Copyright © 2024 All rights reserved ELV.AI

Developed by