On Thursday, Meta announced the launch of an AI model named Meta Motivo, designed to manage the movements of digital agents that simulate human behavior, potentially enhancing user experiences in the Metaverse.

The company has invested billions into AI, augmented reality, and various Metaverse technologies, leading to a forecasted capital expenditure for 2024 projected to reach between $37 billion and $40 billion.

Additionally, Meta is making several of its AI models available for free to developers, believing this open strategy could cultivate the development of improved tools for its platform.

“We think this research can lead to the creation of fully realized agents in the Metaverse, resulting in more realistic non-player characters, democratized character animation, and innovative immersive experiences,” the company stated.

According to Meta, Meta Motivo aims to resolve common issues related to body control in digital avatars, allowing for movements that are more akin to human actions.

Meta is also introducing a new training methodology for language processing called the Large Concept Model (LCM), which aims to separate reasoning from how language is represented.

“The LCM represents a major shift from traditional language models. Rather than simply predicting the next word, the LCM endeavors to foresee the next concept or overarching idea, which is illustrated by complete sentences in a multimodal and multilingual framework,” the company explained.

Other AI tools introduced by Meta include Video Seal, a feature that integrates an imperceptible watermark into videos, making it undetectable without specialized methods but still traceable.