Social App Yubo Introduces Novel Audio Moderation Tool

According to a Pew Research Center study, one in five adults said they would be harassed online for their political views in 2021. Politics has always crystallized differences and debates, but now many of these debates take place on social networks and can sometimes lead to cyberbullying. Politics isn’t the only basis for cyberbullying, as cyberbullying can happen to anyone based on race, gender, ethnicity, appearance, culture or nothing at all.

With the rise in audio exchanges on social platforms, seen with livestreaming on Twitch or  Twitter’s 2020 launch of audio-centric chatrooms, these platforms are facing challenges in their moderation when it comes to audio. It is exceedingly difficult to monitor spoken interaction in real-time, especially when you take into account many different languages, accents and slang. Even Siri can have a hard time understanding us when we are speaking, so a livestream with 10, 20 or even 30 people seems nearly impossible.

Between 2017 and 2021, there has been a six percentincrease in Americans attributing their harassment to their political views. So the problem does not seem to be diminishing over time but becoming even more prevalent. The last campaign for the 2022 mid-terms demonstrated this violence in the debates, and Twitter looks like a battlefield as soon as political news is published.

A recent article by Make Use Of reports a possible solution among social platforms: the launch of an audio moderation system by live social discovery app Yubo. We can already hear the complaints: “I don’t want my conversations listened to” and “what if I’m talking about something private?” Thankfully, with this technology that does not need to be a concern.

In its article, Make Use Of shares the latest announcement from Yubo, a social platform that helps Gen Z make new friends around the world. The platform is known in the industry as a leader in safety innovation and has become the first social app to introduce this type of audio moderation technology in its livestreams. The goal of the technology is to help detect instances of harmful language, hate speech, bullying and threats.

To do so, this new technology records and transcribes 10-second snippets of audio in livestreams of 10 or more people. Then the text is immediately analyzed by an AI, developed in collaboration with Hive – a company specialized in AI moderation of visual, textual and audio content. If offensive, inappropriate, risky or hateful speech is detected, the transcriptionis flagged to Yubo’s human Safety Specialists for review and action if needed.

As with all machine learning systems, this technology is still learning every day to improve its accuracy among four of Yubo’s largest English-speaking markets – the U.S., the UK, Canada and Australia.

What about our freedom of expression, you ask? This moderation tool is not built for censorship, but to help keep the app as respectful and safe a place as possible for Yubo’s users when interacting online. Gone are the days when everyone could be calm and civilized, or at least there will always be those few bad apples that ruin an experience for everyone else. Yubo wants to make sure that adio is left out when it comes to detecting bad behavior, which is why they are adding it to their existing suite of advanced safety tools. Now maybe everyone will keep it kind when discussing politics… just maybe.

Comments are closed.