TL;DR: This week, I watched a few short AI films created by Google’s new AI video tool, named “Flow”. At first its really impressive, making CGI level graphics and film-like scenes for anything you prompt. Watching a bit, you really notice the lack of continuity, believable interaction, and dialogue.
In May, Google released their new AI video tool, called Flow—powered by their Veo 3 generative AI model. Flow can create scenes from text prompts and even generate dialogue. And you have control over details like type of animation, camera actions, and a lot more. Google claims they created Flow in collaboration with filmmakers, but when I looked into those filmmakers I quickly found they were primarily AI “filmmakers” in the first place.
I watched the three short Flow films on Google’s blog, and felt that the films while being visually appealing lacked depth in the characters and interactions needed to make a compelling story.
Flow and the underlying Veo 3 model have a difficult time keeping full consistency between scenes, believable interactions, and dialog between characters.
At its current stage - its not replacing real films, and God forbid it ever does.
Also of note is Google’s Synth ID, which is a hidden watermark that is embedded in each Flow video, so the AI video content can be detected. But Google has sole control of the service, and you must use their service to detect AI video content.
What do you think? Do you think AI filmmakers should call themselves filmmakers?
Solutions:
Support real filmmakers and artists.
Be vigilant in the content you consume
Hire real artists in your creative projects.
Watch this full episode
Follow the #TBOT Show
Get privacy gear
Share this post