In the race for generative AI dominance, Google has taken the lead in the video generation arena with the private preview launch of Veo, its text-to-video model. This move puts pressure on OpenAI, whose competing model, Sora, remains unavailable despite earlier announcements.
While OpenAI showcased Sora's capabilities in February, generating 60-second videos from text prompts, Google has now brought Veo to market through its Vertex AI platform. Google claims Veo can produce high-quality videos in just over a minute, with various visual and cinematic styles available.
Google also highlighted the built-in safeguards in both Veo and Imagen 3 (its text-to-image model) to prevent misuse and inappropriate content generation. However, testing has revealed that these safeguards can be bypassed. Content generated by these models includes an invisible digital watermark created using DeepMind's SynthID technology to identify AI-generated content.
With Veo's launch, OpenAI faces increased pressure to release Sora, which was initially promised for late 2024.
