Google follows up to OpenAI's GPT-4o with new Gemini AI demo video

OpenAI hosted its live stream event "Spring Updates" to show off its freshly-baked AI model GPT-4o and the new ChatGPT app for Mac users. The company showcased the capabilities of its updated AI model through a series of videos.

However, it didn't take long for Google to drop a demo video on X (formerly Twitter), telling the world what a prototype version of its Gemini AI can bring to the table. In the video, Gemini can be seen analyzing the camera video feed in real time and taking questions from a human.

The model responds to questions about what's happening in the video feed and identifies the meaning of the letters written on the screen. It also tries to continue the conversation by asking follow-up questions to the human. Moreover, the conversation between the human and Gemini appears as text on the screen.

The video dropped just a day before the search giant will kick off its annual developer conference, Google I/O 2024. Its keynote will begin at 10 am PT on May 14, where the company is expected to talk about its latest developments around Gemini AI, Android 15, Google Search, and more.

Speaking of updates, a useful feature from Gemini's web version could make its way to the Android app. Android tipster AssembleDebug discovered a "Real-Time Responses" option in the Gemini Settings menu on Android. The feature allows users to read a response as soon as the chatbot starts generating it.

It was reported that Gemini may let you add your preferred music streaming app and access YouTube Music data to offer better search results. Google is also working to add a "Stop" button to Gemini's Android app to let the user halt the response generation mid-way.