Google's New AI Video Model Veo Launches for Businesses

Google has introduced Veo, its latest generative AI video model, now available for businesses to integrate into their content creation processes. Veo was first announced in May, making its debut in a private preview on Google’s Vertex AI platform just three months after OpenAI showcased its own competing product, Sora.

This powerful video model can produce high-quality videos in 1080p resolution, showcasing a variety of visual and cinematic styles based on text or image prompts. Although earlier announcements indicated that the generated clips could last slightly more than a minute, the private preview does not specify any length limitations. The latest example clips shared by Google demonstrate that the videos produced by Veo are impressive, as they are often difficult to distinguish from those created by human filmmakers.

One example that stands out in the Veo clips features a dog, where viewers can notice the realistic consistency of its fur pattern and collar as the animal moves.

The new version of Google’s Imagen 3 text-to-image generator will also be rolled out to all Google Cloud customers through Vertex starting next week. This builds on its initial release back in August via Google’s AI Test Kitchen. Those on Google’s allow list will have access to additional features, such as prompt-based photo editing and the ability to incorporate personal brand elements like logos and styles into the generated images.

However, Veo is not without its flaws. Some instances in the AI-generated videos, like a light shining through a person's hand in a concert clip, indicate that there is still room for improvement.

Both Veo and Imagen 3 are designed with safeguards to reduce the risk of generating harmful content or infringing copyright. Although some concerns about copyright have been identified, Google incorporates DeepMind’s SynthID technology into the content produced, providing an invisible digital watermark. This watermark aims to decrease misinformation and misattribution, similar to Adobe’s Content Credentials system integrated with its generative AI products.

As Veo launches, OpenAI finds itself lagging behind its competitor and faces pressure to deliver its Sora product by the end of 2024. With generative AI content already emerging in advertising, for example, Coca-Cola’s recent holiday campaign has showcased AI-generated visuals. Businesses are motivated to adopt this technology quickly, and according to Google, 86 percent of companies utilizing generative AI report an increase in revenue.

Google, AI, Veo