Eli Collins, vice president of product management at Google DeepMindfirst demonstrated the AI video-generating tool to the company’s board of directors in 2022. While the model was slow, expensive to operate, and sometimes produced unexpected results, he said it was an eye-opening moment for them to see new video clips generated from a random prompt.
Now, just a few years later, Google has announced plans for a tool inside the YouTube app that will let anyone create AI video clips using Veo model of the companyand directly post them as part of the Short Films on YouTube“Looking ahead to 2025, we’ll enable users to create standalone video clips and short films,” said Sarah Ali, senior director of product management at YouTube. “They’ll be able to create six-second videos from an open-ended text prompt.” Ali said the update could help creators who are hunting for footage to complete a video or trying to envision something magical. She insisted that the Veo AI tool isn’t intended to replace creativitybut enhance it.
This isn’t the first time Google has introduced authoring tools to YouTube, though this announcement would be the company’s broadest integration of AI video to date. Over the summer, Google launched an experimental tool, called Dream Screento create AI backgrounds for videos. Before fully rolling out to generated clips next year, Google will update that AI green screen tool with the Veo model sometime in the next few months.
The sprawling tech company has introduced a number of AI video models in recent years, such as Image And Lightbut is trying to coalesce around a more unified vision with the Veo model. “Veo is going to be our model, by the way, going forward,” Collins said. “You shouldn’t expect five more models from us.” Yes, Google may release another video model down the road, but he hopes to focus on Veo for the foreseeable future.
Google faces competition from a number of startups developing their own text-to-video tools. Sora by OpenAI is the most prominent competitor, but its AI video model, announced in early 2024, is still not public and is only available to a small number of testers. As for the tools that are widely available, AI Runway Startup has released multiple versions of its video software, including a new tool for converting original videos into alternate reality clip versions.
YouTube’s announcement comes as AI tools have become increasingly controversial among creators, who sometimes view the current wave of AI as steal from their work and try to undermine the creative process. Ali doesn’t see AI-generated tools coming between creators and the authenticity of their relationship with viewers. “It really comes down to the audience and what they care about—not necessarily the tools,” she says. “But if your audience cares about how you do it, that’s going to be unlocked through the description.” Google plans to watermark every AI-generated video for YouTube Shorts with ID Summarywhich embeds an unrecognizable tag to help identify the video as synthetic, as well as includes a “generated with AI” disclaimer in the description.
Influencers of hustle culture have tried algorithm game by using a variety of third-party tools to automate the creation and monetization process with minimal effort. Will the Veo integration next year result in a flood of low-quality, spammy YouTube Shorts dominating users’ feeds? “I think our experience with recommending the right content to the right viewer is going to work in this AI world at scale, because we’ve done it at this scale,” Ali said. She also pointed out that YouTube Standard Guidelines still applies regardless of what tool is used to create the video.
AI art often has a unique aestheticThis could be a concern for video creators who value individuality and want their content to feel unique. Collins hopes that Google’s fingerprints don’t appear on all AI video output. “I don’t want people to look at this and say, ‘Oh, that’s a DeepMind model,’” he says. Getting the prompt to create AI output that matches what the creator envisions is a core goal, and avoiding overt aesthetics is important for Veo to achieve widespread adaptability.
“A big part of the journey is really building something that’s useful for people, that’s scalable, that’s deployable,” Collins said. “It’s not just a demo. It’s being used in a real product.” He believes that putting the generative AI tools right inside the YouTube app will be transformative for creators, as well as DeepMind. “We’ve never really built a creative product,” he said. “And we’ve certainly never done it at this scale.”
