Google DeepMind is aiming to surpass OpenAI in video generation with its new AI, Veo 2.
Google DeepMind is aiming to surpass OpenAI in video generation with its new AI, Veo 2.
This
next-gen model can create 2-minute videos in up to 4K resolution, four
times the resolution and six times the length of OpenAI's Sora.
Currently, Veo 2 is in an experimental phase in Google’s VideoFX tool,
producing videos at 720p and 8 seconds long.
Veo
2 improves upon its predecessor by offering better physics, camera
controls, sharper textures, and more realistic motion, light, and fluid
dynamics. However, it still faces challenges with character consistency
and detailed motion. DeepMind is working with creators like Donald
Glover and the Weeknd to refine it.
Veo
2 was trained on vast amounts of video data, though there are concerns
about its impact on creative industries. Google is using watermarking
technology to prevent misuse, such as deepfakes. Alongside Veo 2,
DeepMind also announced upgrades to Imagen 3, improving its ability to
generate high-quality, detailed images.