AI chatbots are capable of producing believable, if occasionally overconfident, natural language responses, but what if an AI was able to create unique virtual worlds only from text descriptions?
With its suite of developer tools, which are intended to enable just about anyone to create experiences and virtual goods in the youth-focused virtual ecosystem, Roblox is investigating that prospect.
The business claims to be testing two alternative generative AI implementations, one that can generate code based on text inputs and another that can generate virtual materials based on a natural language prompt (the automobile texturing in the video below). In the upcoming weeks, both will be made available to Roblox Studio developers.
“… Some creators know how to code, but may have limited experience creating high-fidelity 3D models. Others may be more experienced with model design, but less experienced with code,” CTO of Roblox Daniel Sturman published a blog entry. “In both cases, we see a future in which even a beginner (such as myself for example – I’m a terrible artist!) can get a running head start as they look to bring their imagination to life in a Roblox experience.”
The firm has recently made clear that it wants to become the metaverse’s infrastructure, which includes both the toolset used to create virtual worlds and the locations where people go to hang out in them. Roblox has rivals (Epic Games, for example), but AI-based gear like this might make it an even more appealing platform for inexperienced game creators to create the popular virtual worlds that keep young players glued to Roblox.
