How many times have you watched a TikTok or YouTube short and thought “This is fun, but what if there were some strange pandas desperately trying to shove coffee from a morphing teacup into their nose or maw?” Great news.
For all the folks begging for such a feature, YouTube is here for you. For everybody else—especially those concerned about a possible future where cheap, disturbing, and cringe-worthy deepfake videos get streamed from the biggest name-brand platforms—sorry to say the future is here, as much as we wish it weren’t. On Thursday, Google announced that its TikTok clone YouTube Shorts will eventually be getting several new creator features, including an AI video generator called “Dream Screen” built directly into the app.
During its Made on YouTube event, the company’s execs tried to emphasize they want to let anyone “create.” Dream Screen is a tool that lets users generate AI background videos or stills using a simple text prompt. The company did not reveal what model it’s using for these AI videos, but the videos already look similar to the Imagen-based AI-generated clips Google displayed late last year. As CEO Neal Mohan explained, according to TechCrunch, the general idea is to let users’ imaginations run wild. People could pretend they’re in an underwater castle, or flying beside imaginary dragons.
It’s the sort of imaginary tech evangelist babble we’ve come to expect from companies hyping the possibilities of AI, but coming from a giant like Google it’s even more ludicrous than the waning hype around the VR metaverse. As more AI video generators come into the mainstream, the main issue with the generated content is that they’re freaky looking, even at the best of times. While AI image generators have become very capable of producing somewhat believable images, AI video has not yet made that same leap. As Google’s own promo video suggests, the text-to-video software is merely a novelty that’s more disturbing than imaginative.
There’s a few other bits of AI coming to YouTube, including AI-generated dubs in more than 120 languages, as well as AI-recommended video topic, outlines, and audio.
The other announcement from Thursday’s event is much more substantial, at least for the amateur video crowd. Google is planning to release its own, free YouTube Create mobile app that offers some simplified video editing tools on the go. Users can upload video clips from their phones and the app offers some simplified trimming, captioning, voice-over, and audio cleanup tools. There’s also a library of filters, effects, transitions, and royalty-free music. It works in both the horizontal and vertical aspect ratios for regular YouTube Videos as well as Shorts.
Users can download the edited video to their device, and there’s an option to immediately upload the piece to YouTube. The app is currently in beta on Android in eight countries, including the U.S., Germany, France, the UK, Indonesia, India, Korea, and Singapore, according to The Verge. The company does plan to release an iOS version next year. It’s similar to the CapCut app that’s used for making TikTok videos, and it’s nice to have free video-cutting apps that won’t bombard you with ads or lock features behind a paywall.
YouTube execs said the company will start “testing” Dream Screen later this year and it should roll out early next year. That leaves some time for Google to keep working on its text-to-video capabilities, but Google’s focus on being first in the AI rat race has only helped show just how limited, and problematic, generative AI currently is. Microsoft hosted its own product showcase Thursday, and the big headline was that it’s adding its copilot chatbot into Windows 11.
YouTube claimed Shorts currently has over 70 billion daily views—up from 50 billion in January this year—and 2 billion monthly active users. A video editor is certainly handy, but it’s hard to see how disturbing AI video will cut down on the schlock already pervading the TikTok-like platforms.