Roblox has introduced a new AI feature designed to enhance communication by rephrasing inappropriate language in real time. Previously, the platform utilized filters that replaced unsuitable language with hash marks (####), which disrupted chat conversations. The updated system, as explained by Chief Safety Officer Rajiv Bhatia, will substitute profanities with more acceptable alternatives, ensuring clarity in discussions.
When users send inappropriate messages, such as “Hurry TF up,” the AI will modify it to “Hurry up!” while notifying others in the chat about the change. Users will also be informed about the specific language that was altered. Despite the rephrasing, those who repeatedly use foul language may still face penalties for violating Roblox's policies.
This feature is currently available for chats among age-verified users within similar age brackets and supports all languages offered by the game's translation tools. Following concerns about safety, particularly regarding adult interactions with minors, Roblox implemented mandatory age verification in January. However, legal challenges persist, with lawsuits filed by LA County and Louisiana's Attorney General alleging that the platform exposes children to potential dangers.