Google Launches AI Video Tools in YouTube Shorts, Photos
Starting today, Google is introducing generative AI video tools to both YouTube Shorts and Google Photos. While creators will be the first to try these features on YouTube, users of Google Photos in the U.S. will also begin seeing AI video capabilities. Although these features currently rely on the older Veo 2 model, Google plans to upgrade to Veo 3 later this summer.
This rollout follows earlier promises from YouTube CEO Neal Mohan, who said the company would integrate AI tools into Shorts. Previously, creators had access to AI-generated backgrounds. Now, however, they can generate new video elements using only a text prompt. In addition, creators can start with a photo to create short AI-generated videos, adding effects directly from the Shorts camera.
YouTube is also launching an “AI Playground” hub. This section will house all AI tools, feature examples, and offer prompt suggestions. Although these features are initially limited to the U.S., Canada, Australia, and New Zealand, Google plans a broader rollout soon.
Google Photos Joins the AI Push
While YouTube’s tools are built for content creators, Google Photos brings a lighter set of features for everyday users. Beginning today, U.S. users can turn static images into short AI-powered clips. Options are limited for now, offering either “Subtle movements” or a playful “I’m feeling lucky” animation style. Still, it opens the door to more creative uses of existing media.
Within the next few weeks, a “Remix” feature will be added. This will allow users to apply stylistic effects to images quickly, expanding creative possibilities without requiring advanced editing skills. Although only in the U.S. for now, further expansion seems likely.
Photos will also introduce a new “Create” tab, similar in concept to YouTube’s AI Playground. It will collect all generative tools in one place and appear prominently in the app’s navigation bar. This tab is expected to roll out in August.
Watermarking and Safety Still a Focus
As AI tools become more integrated, Google is emphasizing transparency. All AI-generated content will carry a SynthID watermark to help distinguish it from authentic video or photo material. While Veo 2 content may not be hyper-realistic, Veo 3’s improved fidelity has already shown how convincing AI video can become.
Although Google continues its safety analyses to detect misuse, the company admits its models are not foolproof. Trolls and bad actors still pose challenges, but safeguards are evolving alongside the tools.
By embedding AI more deeply into consumer apps, Google is making its generative models more visible and accessible whether users seek it out or not.
