Last night, Google hosted its much-anticipated annual developer conference, and the event was dominated by AI announcements. Google I/O, held in Mountain View, California, offered a glimpse into the future possible with AI. During the event, Google also introduced his two new generative AI tools, Veo and Imagen 3. Veo can be used to generate videos from text prompts (did someone mention his OpenAI's Sora?), and Imagen 3 is called Google's most “sophisticated text conversion tool.” “Image model” is not available yet.
Google's blog post announcing two new AI tools says: “Over the past year, we have made incredible progress in improving the quality of generative media technology. We are working closely with the creative community to explore how generative AI can best support the creative process. , we've worked hard to ensure that our AI is as helpful as possible every step of the way. Today, we're introducing Veo, our newest and most advanced video generation model, and the highest quality text-to-image ever. Introducing Imagen 3, a transformation model released by filmmaker Donald Glover and his creative studio Gilga, as well as artists Wyclef Jean, Marc Rebillet, and songwriter Justin Tranter. The new demo recording was created with the help of our Music AI Sandbox.
Google announces Veo
Veo, which Google says is its most advanced video generation model to date, can produce high-quality 1080p videos in a variety of styles. This AI understands cinematic concepts such as “time-lapse” and “aerial photography” and can produce videos that are true to the filmmaker's vision. The introduction of Veo promises to be a game-changer for video creators by providing tools that not only capture the essence of a prompt, but also maintain realism and consistency across sequences. It can be said that Google is competing with his OpenAI in this field, since recently OpenAI announced the text-video generator Sora.
To demonstrate Veo's capabilities, Google collaborated with famed film director Donald Glover and his creative studio Girga. Their joint project demonstrated the potential of his Veo in professional filmmaking and suggested how AI can become an integral part of the creative process. As Google continues to improve his Veo, selected creators will be able to participate in private Veo previews to explore VideoFX's features.
About Imagen 3
At the same time, Google introduced Imagen 3, which can generate high-quality images from text prompts. Google says the model represents a significant advance, producing photorealistic images with intricate detail and minimal visual distractions. Imagen 3 excels at understanding and interpreting the nuances of text prompts, allowing you to create images that closely match your specifications.
From detailed wildlife portraits to dynamic landscape photography, Imagen 3's capabilities were showcased through a series of generated images that reflected the complexity and style of the specific prompt. The model's enhanced text rendering capabilities also extend its applicability to personalized messaging and professional presentations.
“Imagen 3 is our highest quality text-to-image conversion model. It produces an incredible level of detail and has far fewer distracting visual artifacts than previous models. Generate photorealistic, life-like images. Starting today, Imagen 3 is available to select creators by taking advantage of ImageFX's private preview and joining the waitlist. will be coming soon to Vertex AI,” Google said in a blog post.
