AI video just took a startling leap in realism. Are we doomed?
1 min read
Summary
Veo 3 is Google’s newest AI-powered video generation model, released in Sept 2023 and capable of creating 8-second clips with synchronized audio, a first for the company’s AI tools when it comes to video synthesis.
It generates videos at 720p resolution, starting with either a text description (“prompt”) or a still image, and iteratively refines random noise into a coherent video matched to the description.
Google also launched Flow, an AI filmmaking tool that combines Veo 3 with its Imagen 4 image generator and Gemini language model, allowing creators to describe scenes in natural language, manage characters, locations and visual styles in a web interface.
Both tools are accessible to US users subscribed to Google AI Ultra, costing $250 monthly for 12,500 credits, with Veo 3 videos costing 150 credits per generation.
The growing sophistication of AI-generated media is nearing the point where it is difficult to distinguish such content from authentic media, raising implications for media literacy.
