Roblox chief scientist envisions AI generation of games by voice

Morgan McGuire, who is the chief scientist at Roblox, revealed in the recent Techfirst podcast about incorporating AI into making games using voice only. He also revealed the plan to use AI for real-time translation of voice chat in 45 languages used by users.  

“The content is why people are there…That amazing content, Roblox produces zero of it. So part of our community, and hopefully someday 100% of our community, creates content. It’s a user-generated content platform.” 

– Morgan Mathew in the Tech First Podcast

The content made by people in the game is 3d generation. Although the 3d generation tools in the game are relatively straightforward to use, they still require programming and the knowledge of 3d modeling.

For this purpose, the game has started using conversational AI for players to make avatars and tools. In the future, the platform is planning to use AI in game development, environments, and tools with just voice.

Mathew said, “It’s just up-leveling what you’re doing. Instead of typing every single thing using your mouse, we can have AI take your intent and produce it for you.”

Roblox will use AI in real-time voice translation  

In the podcast, Morgan also said, “We want to ultimately do voice fonts … so basically voice role-playing kind of transitions. So I’m speaking in my voice, but you’re hearing a dragon. So it fits my character, right?”

The chief scientist also revealed that they are planning to use AI in real-time communication as well. The AI will translate 45 languages in real-time voice translation for stronger community building in Roblox. This feature will be known as “voice fonts”

A lot of Roblox players reside outside of the home base of the game, which requires cross-cultural communication within the game. Mcquire revealed the game already uses AI to monitor the real-time voice chat in 45 different languages, in total 50,000 messages a second. He says the results with the use of AI are far better than any human could have achieved.