Nowadays many of your tasks have become easier with the help of AI. You can complete your work from photo editing to video editing in minutes. But have you ever imagined that you are in a digital world where you can hear the sound of the wind blowing through the trees, the waves of the ocean and even a character standing in front of you responding to your words. Not only that, just by giving a command you can change the weather there, rotate the camera angle or add a new person or object of your choice. It may sound like a video game, but Google DeepMind’s new AI model Genie 4 claims to work in that direction. According to the company, it can create such a digital world in real-time, which will change according to your instructions.

Google on Tuesday introduced the Gen 4 as its new global model. Currently, it is in early research preview and is being tested with some trusted partners.

What is Google DeepMind’s Genie 4?

Genie 4 is designed to create an interactive environment rather than just showing a picture or pre-recorded video. According to Google, this model utilizes various capabilities of the Gemini family. Gemini Omni helps create simulations that look and act like the real world, while Gemini’s other models contribute to ensuring events remain relevant and contextual in the digital world.

Its specialty is that the user can modify the simulation even while running. For example, you can make it rain, move the camera in a different direction, or add a new character to the world with a text command.

Now there will be sound in the digital world

One of the most interesting features of Genie 4 is its audio generating feature. Google claims it can even generate sound in real-time. It includes wind blowing through trees, sound of ocean waves, background music.

That is, not only objects can be seen in the simulation, but sounds can also be heard according to that environment. According to the company, the model can also communicate between characters who respond to user instructions or events that occur in the simulation. For this, there will be no need to pre-record each voice, as the conversation can be tailored to the changing environment.

How will Genie 4 work with Gemini?

According to Google, Gemini 3.8 Flash acts as a guide during the simulation. He constantly looks at the digital scene, breaks down the big goal into smaller tasks and checks that each step has been completed correctly. Apart from this, Google DeepMind is also experimenting with Gemini 4 Argon and Genie 4 in its Gemini robotics research.

Also focus on safety

Genie 4 is still in early research stages, so it is not guaranteed to provide accurate results every time. The appearance of some objects may change if the simulation runs for a long time or if changes are made frequently. Even when progressing through a scene, small errors may appear. Google says that the model has undergone internal security checks and security measures have been incorporated to prevent harmful content. Additionally, the company’s SynthID technology will be used in video and audio produced from the Genie 4. It adds an invisible identifier to content, helping identify whether it was created by AI.

What will be special at Apple’s event today? What new products can be unveiled, you will get many surprises