Google explores new audio, video, and image generation AI

The Google I/O developer conference showcased a ton of new features coming to users now and in the coming weeks and months. This includes updates to how you can create videos, images, and music using generative AI.

The Google team has been working on improving its generative AI, and among its advances (like some pretty cool image search AI) are some new video creation features using an agent called Veo. Imagen 3 text-to-image updates and some new features in Google's Music AI Sandbox.

First, Veo is a new generative AI agent that can create 1080p videos using text, images, or voice prompts. Veo matches the videos you can create to the style of the photos being used, and takes advantage of several new tools to help you build the video you envision. For example, Veo now understands terms like “time-lapse,” “tracking shot,” and “aerial shot” to better create the frames you need.

“Using Veo, we have improved how models learn to understand video content, render high-resolution images, simulate the physics of the world, and more. “It's an even more useful product that accelerates advances across research and helps people interact and communicate in new ways,” said Eli Collins, vice president of product management, and Douglas Eck, senior research director. stated in Google's post announcing the feature.

Imagen 3 is the latest image module that Google announced during the conference. Google says it “better understands the natural language and intent behind prompts, and incorporates finer details in long prompts.” The result is a more realistic and better image with more detail than previous models. Imagen 3 now handles text art better than previous versions, allowing you to create personalized messages and other text decorations.

Finally, Google has spent time on the Music AI Sandbox, which allows musicians to create music using AI. New additions include the addition of musical instruments and more ways to express your creative style. Google has partnered with musicians like Wyclef Jean, Marc Rebillit, and Justin Tranter to continue developing the Music AI Sandbox and expand Gemini AI's music creation capabilities.

All of these features are in testing. Veo and Imagen 3 are available to select creators through VideoFX (if he is not already on VideoFX, he must be on the waiting list). Meanwhile, the Music AI Sandbox is still limited to artists that Google has asked to participate.