The centerpiece of YouTube's Made On event on Wednesday morning was, you guessed it, artificial intelligence: The company announced that it would be integrating Google DeepMind's AI video generation model, Veo, into YouTube Shorts to let creators generate high-quality backgrounds and six-second clips.
Veo was announced at Google's I/O 2024 developer conference as a state-of-the-art video generation model. The technology competes directly with OpenAI's Sora and other competing video generation models such as Pika, Runway, and Irreverent Labs. It can create a variety of cinematic-style 1080p video clips.
Veo in Shorts is set to be a major upgrade to YouTube's AI-powered “Dream Screen” feature, which launched in 2023 and allowed creators to generate backgrounds for their Shorts using text prompts. YouTube believes the Veo model will further enhance the video background generation process, helping creators produce more striking clips. One of Veo's main benefits is the ability to edit and remix previously generated footage.
Additionally, for the first time, creators can create six-second standalone video clips for their Shorts. Once creators select “Create” and complete the prompts, Dream Screen will generate four images for them, after which they can select the photos and turn them into a video.
This new feature allows creators to add transitional scenes to their videos to create smoother transitions and tie together an overall story. For example, a creator could include a scene such as the New York City skyline at the beginning of a tourism video to add more context.
The company plans to integrate Veo into its Dream Screen later this year. Shorts productions will be watermarked using DeepMind's SynthID technology to indicate they are AI-generated.
Image credit: YouTube
In addition to integrating with Veo, the company also announced that YouTube will be adding a number of new features, including “jewels” or gifts, digital items that viewers can send during live streams. The feature appears to be similar to TikTok's “gifts.” Jewels aims to give viewers a new way to interact with creators and actively participate in live streams. The feature will begin rolling out to the live streaming vertical in the US.
YouTube has also expanded its auto-dubbing tool to support more languages, including French, Italian, Portuguese and Spanish. Most notably, it is testing “expressive speech” — a feature that transfers the creator's tone, intonation and ambient sounds into the dubbed audio to create a more natural experience.
The company is expanding its community hub to more channels, allowing creators and followers to interact with each other by sharing posts and replying to each other.
YouTube is also rolling out the “Hype” feature to other markets. YouTube is initially testing the Hype tool in Brazil, Turkey and Taiwan, allowing users to show their support for their favorite creators. Videos with the most Hype points will be featured on a special leaderboard.
Additionally, the company revealed at today's event that creators can now use AI to help them brainstorm video ideas within YouTube Studio, as well as create AI-generated thumbnails and reply to followers with new AI-assisted comments.