YouTube Shorts has upgraded its Dream Screen AI background generator to incorporate Google DeepMind’s latest video model, Veo 2, which will also generate standalone video clips that users can post to Shorts. “Need a specific scene but don’t have the right footage? Want to turn your imagination into reality and tell a unique story? Simply use a text prompt to generate a video clip that fits perfectly into your narrative, or create a whole new world,” coaxes YouTube, which seems to be trying out “Dream Screen” branding as an umbrella for its genAI efforts.

In a blog post announcing the incorporation of its parent company’s new tech, YouTube hovers between promoting Veo 2, a multimodal LLM, and touting the more teen-friendly Dream Screen handle.

“Veo 2 understands real-world physics and human movement better, making its output more detailed and realistic,” YouTube says, listing the ability to swap lenses and conjure various visual styles and cinematic effects among Dream Screen’s new features.

“Together with Dream Screen, Veo 2 generates state-of-the-art, high-quality videos in a wide range of subjects” while also making “improvements so Dream Screen now generates videos faster than before,” YouTube adds.

“To access clip generation, Dream Screen users need to open the Shorts camera, open the media picker, and tap the ‘Create’ option at the top,” The Verge instructs. “You can then enter a text prompt describing what you want to generate and select your desired style, lens, cinematic effect, and video length.”

The concept is “interesting,” notes The Verge, “given Google’s latest Veo model is still in early access and only available to the public via a waitlist.”

 Social Media Today writes that “Google’s Veo 2 model generates realistic looking video content, in high definition (up to 4K), and can produce some impressive results based on just text prompts.”

YouTube is adding safeguards to try to ensure everyone knows the Shorts generative videos were made using AI. “They will use DeepMind’s SynthID tool, which watermarks and identifies AI-generated content,” reports TechCrunch.

But while the generative videos “will be clearly marked as being made with AI,” TechCrunch warns that “these labels don’t completely mitigate the potential harm of sharing synthetic content that can mislead viewers.”