# **Google’s Project Astra: Introducing Real-Time Video AI to Gemini Live**
Google is poised to unveil an innovative update to its Gemini AI platform, incorporating real-time video and screen-sharing functionalities into its conversational AI assistant. This capability, first highlighted as **Project Astra**, will soon be accessible through **Gemini Live**, enabling users to engage with AI in a more active and user-friendly manner.
## **What Is Project Astra?**
Project Astra was initially presented at **Google I/O 2024**, where it demonstrated the potential to process and respond to live video input. In contrast to conventional AI assistants that depend solely on text or static images, Astra showcased how AI could **analyze live video, identify objects, and even recall previous interactions**.
For instance, during Google’s demonstration, Astra was capable of:
– Recognizing **code on a computer screen** and offering explanations.
– Clarifying **how speakers function** by examining their design.
– Identifying a **network diagram on a whiteboard** and detailing its elements.
– Recalling where a user had placed their **glasses** earlier in the video.
This type of engagement signifies a considerable advancement in AI functionality, transforming it into a **real-time assistant** rather than merely a chatbot.
## **Gemini Live: The Next Evolution for Google’s AI**
At **Mobile World Congress 2024**, Google announced that this Astra-enhanced feature is prepared for public launch. The update is set to be deployed to the **Gemini app on Android** later this month, permitting users to:
– **Utilize their phone camera** to supply a live video feed to Gemini Live.
– **Share their screen** instantly and inquire about what is being displayed.
This allows users to simply **show** Gemini something instead of describing it, making interactions more rapid and intuitive.
## **How Will It Operate?**
Presently, Google’s multimodal AI can handle **text, images, and documents**, but its video processing has been erratic. Occasionally it can summarize YouTube videos, while at other times it may falter. With this update, Gemini Live will be equipped to:
– **Examine live video feeds** from a smartphone camera.
– **Deliver real-time responses** based on its observations.
– **Assist with tasks such as troubleshooting, shopping, and navigation**.
For example, when shopping online, you could share your screen with Gemini Live and request outfit suggestions. Alternatively, if you’re discovering a new area, you could use the camera to inquire about landmarks or directions.
## **Subscription and Accessibility**
The new video functionalities will be included in **Gemini Advanced**, which necessitates a **$20 monthly AI Premium plan**. This subscription also provides access to Google’s most sophisticated AI models.
Nonetheless, executing real-time video processing requires substantial computational resources, indicating that Google will likely **incur losses per user**—at least initially. Regardless, the company is advancing, aiming to contend with OpenAI and other leaders in AI.
## **The Future of AI Assistants**
With this update, Google is positioning Gemini as a **genuine AI assistant**, capable of **real-time, multimodal interactions**. While it remains uncertain whether the public version will live up to the striking Astra demonstration, this represents a significant move towards creating AI that is more **interactive, beneficial, and human-like**.
As AI progresses, features like **live video processing** may revolutionize our engagement with technology, leading to digital assistants that are more **intelligent, responsive, and applicable** in everyday scenarios.
