It's an audio only chatting app where all conversations happen in real time.

Clubhouse was launched in March 2020 but the invite only APP already has 10 million weekly active users.

Its executives wanted to host everything from political rallies to company meetings.

But with fame comes scrutiny.

Critics are questioning how the APP will handle problematic content such as hate, speech, harassment and misinformation.

Reuters social media correspondent to Elizabeth Clifford Content Moderation at scale is something that even the biggest social media platforms grapple with and moderating any kind of live content.

Whether that's video audio, that is difficult because it is immediate.

It's evolving.

Unless the content is recorded, then it doesn't live as a static piece that your answer could just be to take it down.

Um, so you're enforcement Actions also might be different in how you handle that and then moderating live audio in particular, um, is also difficult because it's more cumbersome and time consuming, um, than detecting and analyzing text, whether that's with machine learning tools or with human moderators who may take more time to, um, listen and get the nuances of audio than reading text.

It's a challenge being faced by a raft of social platforms from Twitter's fledgling live audio feature, called Spaces to Discord, a video game centric platform with live voice chat.

Facebook is also reportedly dabbling with an offering.

Companies vary when it comes to handling violations.

Clubhouse, which is pretty new, Um, says that it does record chats but that it deletes the chat if they haven't had a user report by the end of that live session.

So you can report something after the fact that they may not have the audio to investigate.

Um, that, though it says it will still take action.

Um, discord doesn't really doesn't record at all.

So it relies on metadata on other context untrusted users reporting what happened.

Whereas Twitter Spaces, which is Twitter's new live audio feature within the site that's still being tested with a small number of users, they say that they'll keep the audio for 30 days 19,000 incident Um, and they also say that they're looking into doing more proactive detection on the site, though it's not clear how that will work.

Yet the challenges of live audio are set against a broader content moderation battle where platforms can find themselves in the firing line for being too restrictive or dangerously permissive.

We want to do more here, uh, to prohibit the kind of divisive and inflammatory rhetoric that has been used to sew discord.

But audio is clearly a battle worth fighting.

Getting it right could help spark new waves of business monetizing not only users eyeballs, but their ears, too.