Google has introduced Veo 2, its latest AI video generator, saying it creates more lifelike videos than OpenAI’s tool, Sora.Google Veo 2 is designed to understand real-world physics better and capture human movements and expressions more accurately. 

Google Veo 2 is available through Google Labs’ VideoFX platform, but you’ll need to join a waitlist using a Google Form to get access. Google also claims Veo 2 can handle filmmaking styles, allowing users to choose genres, lenses, and effects. It even supports 4K resolution.

For now, the old Veo stays on Vertex AI. Any video made with the Veo 2 will include a Google SynthID watermark to show that it’s AI-generated. While Veo 2 has improved, it can still occasionally create odd mistakes, like extra fingers, though these are now less common.

Google Veo 2 Competing with Sora

Veo 2 is designed to compete with OpenAI’s Sora, which became available to paid users recently. Google claims that in its internal tests, audiences preferred videos made with Veo 2 over those from Sora.

Google first announced Veo in May at its Google I/O event, showcasing its abilities with a video featuring Donald Glover (Childish Gambino).

The Bigger Picture in AI Video Creation

While the technology continues to improve, challenges remain. Other,  Smaller companies like RunwayML, Pika Labs, and Luma AI have also been innovating in this space:

• RunwayML: Advanced controls have recently been added to its latest Gen-3 Alpha Turbo model.
• Pika Labs: An update was released that allowed users to add their own characters to the video.
• Luma AI: partnered with AWS to make its tools available for enterprise use.

Still, people are wary about AI videos, as certain tools, including Sora, often generate unrealistic figures or inconsistent results. Even the recent showcases, like the one at the Game Awards, have left viewers in doubt about AI quality.

That said, a few filmmakers see the potential. Director James Cameron joined Stability AI’s board, and actor Andy Serkis is creating an AI-focused production company. On YouTube, creators are already using Veo’s tools to make backgrounds for Shorts, saving time.

Updates to Imagen 3

Google also updated its image model, Imagen 3, to make photos more vibrant and realistic. The tool can now handle a variety of art styles, such as photorealism, anime, and impressionism, and is better at following user instructions. Imagen 3 is accessible through Google’s ImageFX.

Both Veo 2 and Imagen 3 aim to make creating content faster, easier, and more creative for users.

// Did you find this information helpful? Let us know in the comments!