Roblox has launched an AI-powered feature that can reprogram inappropriate language in real-time. The online game has been using AI filters to block any language that goes against its policy for some time now, but it is replacing the censored chat with a series of hash signs (####). Roblox acknowledges that encountering too many hashmarks can be disruptive and make conversations hard to follow. This new feature will replace words and phrases with what the AI deems more appropriate alternatives.
Rajeev Bhatia, Chief Safety Office of Roblox, said that the game started with abuses. For example, if a user sends “Hurry up TF up” in chat, the system will interpret it as “Hurry up!” Will replace it with. When a message is rewritten, everyone in the chat will see a note, and the sender will see in which language it was edited. A user who keeps swearing in chat will be penalized for breaking Roblox policy, even if the AI rewrites their messages. “As these systems scale, they create a flywheel for civility, where real-time feedback helps users learn and adopt our community standards,” Bhatia said in a blog post.
Refrasing has been introduced for chats between age-verified users in the same age groups and in all languages supported by the game’s translation tool. Roblox introduced a mandatory age verification system in January after reports emerged that it had a “pedophile problem”, with adult players allegedly using the game to groom children. Children under 13 can no longer use in-game chat except in certain experiences, while everyone else can chat with players around their age. However, the age check hasn’t stopped authorities from prosecuting Roblox: LA County said in a lawsuit filed in February that Roblox knew its platform “makes children easy targets for pedophiles.” Louisiana’s AG also recently filed a lawsuit, saying that Roblox “created a public park and filled it with sexual predators who prey on children.”
<a href