Roblox launches real-time AI chat rephrasing to filter out banned language

Roblox has unveiled a groundbreaking feature that utilizes artificial intelligence to rephrase chat messages in real time, transforming inappropriate language into more respectful alternatives. Announced on Thursday, this innovative tool aims to foster a more civil communication environment among users. Previously, the platform's chat filter replaced banned words with hash marks, leading to conversations filled with '####,' which often disrupted the flow of dialogue. With the new feature, messages that would have been obscured will now be rephrased to convey the original intent respectfully. For instance, a message like 'Hurry TF up!' will be modified to simply 'Hurry up!' This change not only clarifies the conversation but also ensures that all participants are informed that a message has been adjusted, promoting a more respectful and engaging chat experience. Rajiv Bhatia, Roblox's Vice President of User and Discovery Product, emphasized the significance of chat in the gaming experience, stating, 'Real-time rephrasing helps keep gameplay and conversations on track while guiding language toward what’s appropriate.' This new feature aims to reduce chat disruptions while upholding community standards. While the rephrasing tool enhances communication, Roblox assures users that its existing safety systems will continue to address more serious behavioral issues. Additionally, the feature supports all languages currently offered through Roblox's automatic translation tools. The company is also enhancing its text-filtering technology to better identify variations of prohibited language, reporting early success in recognizing leetspeak and other attempts to bypass filters, resulting in a twentyfold decrease in false negatives related to sharing personal information. This announcement follows Roblox's recent implementation of mandatory facial verification for users wishing to access chats. This measure comes in response to legal actions from the attorneys general of several states, including Texas and Kentucky, which raised concerns about child safety on the platform amid reports of exposure to grooming and explicit content.