The internet can be a hostile place, and if you’re on a Discord chat, dialogue may be unsavory if you’re in the wrong crowd. If you have a Discord server and want to foster a welcoming atmosphere, you will need help in preventing bad actors from interacting, especially if there are kids on your servers. Enter Kidas’ ProtectMe Bot, an AI-powered software that aims to safeguard young and old Discord users from predators, stalkers, scammers, and other threats.
Kidas is a technology solutions company working on helping Discord servers from turning into the Wild West thanks to its software. ProtectMe Bot utilizes its unique Discord voice chat monitoring technology, marking a first to feature this level of protection. It makes monitoring easy for server administrators and managers thanks to the voice and text algorithms that track privacy violations, cyberbullying, financial scams and more.
“So I’m pretty sure that you’re aware of the toxicity in Discord and scams, harassments, all of those things that are happening in several of the servers,” Kidas CEO Ron Kerbs tells us. “It mainly happens in voice chat, because text chat is fairly moderated. And then we, ourselves, have encountered that, so we decided to create a solution for that specifically for Discord. We come from the gaming space; we have a solution for games. And we got some inbounds from Discord servers and communities to see if they can use our solution. And so we created a Discord bot to address this solution… to address the problem specifically for Discord servers.”
“And it matters for those servers for those two specific reasons,” he continued. “First of all, they’re losing community members, because if you were bullied, harassed, or scammed in a server, you’re probably not going to interact in the server. Maybe you’ll even leave the server because of that one time. And secondly, it saves them a lot of time.”
One of their customers has 17 moderators, joining and logging into the voice chat, etc. With ProtectMe Bot, it helps automate their voice chat and helps them saves a lot of time, money and effort from having to manually moderate. The tool can join and analyze the conversation and analyze the context and content of the conversation. It can even detect if someone is just making loud noises and irritating people in the chat.
“We send alerts to Discord server admins to let them know if something like that is happening,” Kerbs explained. “So they can ban those users or maybe mute them and take action to protect their community members.”
Once ProtectMe Bot notices trouble in a voice chat, a transcript is sent to the admins and managers, where they can use their judgment to decide the fate of the disruptive user.
There’s also a feature to detect kids if you want to prevent them from joining a mature channel. One way the program detects that is the sound of their voice. But if someone is using a voice changer, it can be bypassed, and that’s why there is a second solution via the type of words that are being used.
“If you’re a kid, you’re probably not using the same type of words and frequency of words as the college graduate,” he said. “Based on the type of words that you’re using and the sophistication of the words that you use, we can estimate your age. And that’s basically how we can detect it. So if you’re on the border, it’s very hard to differentiate between someone who’s 16 and someone who’s 18, but you can definitely differentiate between someone who’s 18 and someone who’s 12. Or someone who’s 16 and someone who’s 25. They use different types of words and different types of slang.”
There are ways to trick the system. For example, someone can ask ChatGPT to give them a transcript to make them sound like they’re 25. But in a voice chat where people are usually just chatting normally, it’s a different story.
“If they’re talking normally based on the type of words that they’re using and based on their voice, you can estimate pretty accurately their age,” Kerbs said.
ProtectMe Bot is available in different pricing plans, and you can get it at ProtectMe Bot’s site.
