YouTube is starting to implement GenAI into its Shorts creation hub, with new features like photo-to-video generation and an arsenal of AI-powered video effects powered by the company’s second-best generation model. Not to mention, YouTube’s new AI Playground adds full-on content generation.
On the heels of releasing Veo 3 and discussing at length how powerful AI generation is becoming, Google is officially rolling out new GenAI creation tools for users via YouTube Shorts, with a few adding some utility to captured content. The new features will live in the Shorts creation hub, allowing users to generate new content from still photos, existing videos, or even drawings.
YouTube announced that three major GenAI functions are available to Shorts creators. The first is a photo-to-video feature, which only requires a single image. From that, Google’s GenAI model creates a quick clip based on a suggestion for the model to work with. Google hasn’t detailed the length limit for generated clips, but they’ll likely be rather short.
Shorts is also adding a couple of options in the effects set in the app. In the AI section, two new effects can be selected before recording a video. The first is a fun underwater effect. During the video, the underwater effect will begin to generate frames between organically captured ones to depict the subject in water. Google’s own sample video looks impressive.
The effects section also allows users to doodle drawings on their phone, turning them into more colorful generated images. The latter feature feels a little more out of place than the rest in the YouTube app, but it’s fun nonetheless.
This GenAI update brings Google’s AI Playground into play with YouTube Shorts, as well. The playground is aimed at content generation, rather than adding effects to videos and drawings. Here, users can use prompts like with any other LLM model and create videos and music. That content can then be posted to YouTube Shorts.
It’s unclear whether or not these generated videos are based on user-created prompts or just YouTube suggestions. According to YouTube, generated content will be clearly labeled to indicate the use of AI.
This is more along the lines of what was initially alluded to during a quick announcement last month. It was noted that Google would soon integrate content generation with YouTube utilizing Veo 3, the company’s latest generation model for video. Interestingly enough, the AI tools YouTube announced today for Shorts are powered by Veo 2, but the company notes that they’ll soon run on Veo 3. There is no timeframe set.
YouTube also noted that these features, rolling out now and over the next couple of weeks, would be available in the US, Canada, Australia, and New Zealand.
FTC: We use income earning auto affiliate links. More.
Comments