Key Takeaways
- Google’s AI video generator, Veo, is now in private preview for Google Cloud customers.
- Veo is being marketed to advertisers for making cheaper, quicker ads.
- Imagen 3, Google’s upgraded text-to-image generator, will also be available to all Google Cloud customers next week.
In 2023, AI image generators like Midjourney and Stable Diffusion were impressive, but they became a thing of the past when OpenAI introduced Sora, its AI video generator. The startup unveiled Sora in February 2024 but hasn’t shared any meaningful updates on its release timeline since then. Google, however, has already made its AI video generator Veo, available to customers, despite unveiling it three months after OpenAI at its developer conference in June.
Google announced on Tuesday that Veo, its AI tool for making short video clips from images and prompts, is now available in private preview for customers using Vertex AI, Google Cloud’s AI platform (via The Verge). Veo can churn out 1080p videos in different visual and cinematic styles based on text or image prompts. It’s not clear how long the videos can be, though. At launch, Google hinted they’d be about a minute long, but it didn’t mention any length limits for this preview release.
Google is pitching Veo to advertisers as a way to make ads faster and cheaper. Companies like Mondelez International (the owner of Oreo), WPP, and travel agency Agoda are already using Veo to create marketing content. Google even shared an example of how Agoda’s team used Veo to make a promotional video, and… well, it’s not quite there yet. The videos look semi-realistic, which might work for some advertisers, but anyone serious about their marketing game might want to hold off on AI video generators for now.
The latest version of Google’s Imagen 3 is also here
Google announced that the latest version of its Imagen 3 text-to-image generator will also be available to all Google Cloud customers via Vertex starting next week. This expands on its original US-only release through Google’s AI Test Kitchen in August. Users on Google’s allow list will also get access to new features like prompt-based photo editing and the ability to add their own brand, style, logo, or product features to the generated images.
Google says Imagen 3 creates the most realistic and high-quality images from simple text prompts, with better detail, lighting, and fewer artifacts than previous versions. The images look pretty realistic at first glance, but if you look closely, you’ll see there’s still some work to be done.
Related
Google DeepMind’s AI watermarking tool goes open source
Fighting to separate human from machine
