Roblox has announced a new real-time, AI-powered chat rephrasing feature. This system automatically replaces banned words with more respectful language, moving beyond the platform’s current text filter that simply replaces phrases with the “#” symbol. The company stated that when users encounter strings of these symbols, conversations can become disrupted and hard to follow. Now, instead of displaying hash marks, filtered text will be rephrased into language that is close to the user’s original intent.
For example, a message reading “Hurry TF up!” would previously have appeared as “####” but will now be rephrased to “Hurry up!” Everyone in the chat is notified that the message has been rephrased to maintain civil conversation.
A company executive explained that chat is central to how people connect and play on Roblox. Real-time rephrasing helps keep gameplay and conversations on track while guiding language toward what is appropriate. This approach aims to reduce friction in chat while upholding community standards.
Roblox notes that while rephrasing reduces some disruption, its broader safety system remains in effect for more serious behavior. The new feature is supported in all languages currently available through Roblox’s automatic translation tools.
The company is also upgrading its text-filtering system to detect more variations of banned language. Early results show the system is improving at detecting leetspeak and more sophisticated attempts to bypass filters. This has allowed Roblox to reduce the prevalence of false negatives when users share or solicit personal information by twenty times.
This announcement follows shortly after Roblox introduced mandatory facial verification for access to chats on its platform. That move came in response to a wave of lawsuits over child safety from several state attorneys general. The lawsuits were filed following reports that Roblox was exposing young users to dangerous risks, such as grooming and explicit content.