Meta launches AI models to pave way for ‘fully embodied agents’

Social media and tech giant Meta has launched Meta Motivo, a foundation model for controlling the behaviour of embodied virtual agents, and Meta Video Seal, an open-source model for video watermarking.

Meta Motivo is trained with an algorithm that uses a set of unlabelled motion data used by the AI model to learn human-like behaviour.

Meta stated that advances like this could pave the way for “fully embodied agents” in the Metaverse, leading to more realistic NPCs, widespread character animation, and new types of immersive experiences.

The company says it has recently been focusing on creating more capable agents, with improved robustness and security, as well as developing architectural innovations to further improve its artificial intelligence (AI) models.

According to Meta, its AI model can solve a wide range of full-body control tasks, including motion tracking, pose achievement and reward optimisation, without any additional training or planning.

The model aims to exhibit more human-like behaviour, with Meta adding it already shows an improved level of robustness to changes in the environment, such as gravity, wind or direct perturbations, despite not having been trained for these situations.

Meta Video Seal is a a model capable of adding a watermark with an optional hidden message to videos. The watermark is undetectable to the naked eye and can be discovered later to determine the origin of a video.

According to Meta, the watermark has proven resistant to common video editing operations, such as blurring or cropping, and compression algorithms commonly used for sharing content online. Meta said that it is releasing the Video Seal model under a permissive licence, along with a research paper, training code and inference code, and that a demo is also available.

Meta emphasised the importance of implementing safeguards to mitigate the risks of AI tools such as imitation, manipulation and other forms of corruption, adding that post-hoc watermarking represents a step towards improving transparency.

“We aim to democratise access to state-of-the-art technologies that transform our interaction with the physical world, which is why we're committed to fostering a collaborative and open ecosystem that accelerates progress and discovery,” Meta said in a statement.



Share Story:

Recent Stories


Bringing Teams to the table – Adding value by integrating Microsoft Teams with business applications
A decade ago, the idea of digital collaboration started and ended with sending documents over email. Some organisations would have portals for sharing content or simplistic IM apps, but the ways that we communicated online were still largely primitive.

Automating CX: How are businesses using AI to meet customer expectations?
Virtual agents are set to supplant the traditional chatbot and their use cases are evolving at pace, with many organisations deploying new AI technologies to meet rising customer demand for self-service and real-time interactions.