Google Has Introduced A "real-time Video" Feature In Gemini

Google has finally introduced a "real-time video" feature to its generative artificial intelligence (AI) service, allowing users to interact with Gemini through their smartphone cameras for question-and-answer sessions. 

According to a report by The Verge on Monday, this new interaction feature enables the AI to view the user's screen via the smartphone camera and respond to inquiries related to the images captured by the camera in real time. 

This was also confirmed by Google spokesperson Alex Joseph in an email sent to The Verge. The feature arrives nearly a year after Google initiated Project Astra, a research endeavor aimed at enhancing AI assistants' understanding of their users' surroundings. 

The discovery of this feature in the commercial service was first made by a Xiaomi phone user who accessed Gemini and shared their experience on the Reddit community. 

Shortly thereafter, other users uploaded similar videos demonstrating the use of the Gemini feature, which is a result of Project Astra, showcasing Gemini's ability to read text from their phone screens. 

The capability to read from the screen is one of two features that Google promised in early March 2025 during the Mobile World Congress (MWC), as part of the rollout to select Gemini Advanced customers. 

As a result, users utilizing the "Real-Time Video" feature from Gemini can now display live visuals from their smartphone cameras and request the AI assistant to answer questions related to those visuals. 

For a demonstration of how this feature works, users can refer to the video released by Google in early March 2025, coinciding with the introduction of this feature at MWC. 

Tag: Artificial Intelligence Gemini Smartphone Cameras