Nvidia unveiled a prototype AI avatar that lives on your PC's desktop at CES 2025. AI Assistant; R2XIt looks like a video game character and it can help you navigate apps on your computer.
The R2X avatar is rendered and animated using Nvidia's AI models, and users can use their avatar in popular LLMs such as OpenAI's GPT-4o or xAI's Grok. Users can chat with R2X by text and voice; You can upload files to it for processing or enable the AI assistant to see what's happening directly on your screen or camera.
Tech companies have been creating a lot of AI avatars lately, not only in video games, but also for business and consumers. Early renderings are strange, but some think these avatars are a potential user interface for an AI assistant. With R2X, Nvidia is trying to combine next-generation video game skills with cutting-edge LLMs to create an AI assistant that feels like a human.
The company plans to open-source these avatars in the first half of 2025. Nvidia sees this as a new user interface to build with developers, allowing users to plug in their favorite AI software products or even run these avatars locally.
love it. Microsoft's Recall feature (It was Delayed due to personal concerns.) R2X can take continuous screenshots of your screen and work with an AI model to do this, although this feature is disabled by default. When enabled, It can provide suggestions about the applications running on your computer and help you perform complex coding tasks, for example.
R2X is still a prototype, and even Nvidia admits there are still some bugs. In demos with TechCrunch; Nvidia's avatar had an uncanny valley feel to it — its face was sometimes stuck in odd positions, and its tone felt a bit harsh at times. In general, I find it a little strange to have an inhuman avatar staring at me while I work.
R2X generally provides helpful instructions and looks accurately at what's on the screen. But at one point the avatar gave us wrong instructions and later the avatar couldn't see the screen at all. This may be a problem with the underlying AI model (in this case, GPT-4o), but the example shows the limitations of this early technology.
In a demonstration, An Nvidia product lead showed how R2X can help users view apps on your screen. in particular, R2X helped to use Adobe Photoshop's next-generation fill feature. The photo we chose is of Nvidia CEO Jensen Huang standing in an Asian restaurant with two restaurant employees. Nvidia's clone was an illusion and gave incorrect instructions on where to find the next generation fill feature in Photoshop. Later it lost the ability to see the screen, but after switching the AI model we used to xAI's Grok. The statue regained its screen-viewing ability.
In another demonstration, R2X was able to answer questions about it after loading a PDF from the desktop. This process is powered by a local retrieval augmented generation (RAG) feature, which allows these AI clones to pull information from a document and process it using the underlying LLM.
Nvidia is using some AI models from its video game division to power the look of those images. Nvidia uses its RTX neural faces algorithm to generate avatars. face Nvidia is using a new model to automate lip and tongue movement. Audio2Face™-3D. The model seems to be stuck in some places, and the avatars are held in awkward positions.
These R2X avatars will be able to act as personal assistants in Microsoft Teams meetings, the company said.
An Nvidia product lead said the company is working to give these AI avatars agent capabilities, and that R2X could one day run on your desktop. These capabilities seem a long way off, and they will need to collaborate with software makers like Microsoft and Adobe, who are trying to create similar agent systems themselves.
It was not immediately clear how Nvidia is generating sounds in these products. R2X's voice when using GPT-4o is unique from ChatGPT's preset voices, and xAI's Grok chatbot has no voice mode at all.