
YouTube added a feature called Make Me Move to its Shorts AI playground in early 2026. Instead of generating video from a text prompt, it takes a single still photo of a person and animates that person into a Short based on a preset reference movement. The result is a short clip where the subject appears to perform a physical action they never actually did.
Make Me Move requires two inputs. The first is a full-body photo of one person, selected from your camera roll. The second is a reference movement, chosen from a set of presets inside the Shorts creation screen. YouTube currently offers a handful of movement templates, including a karate combo. You access the feature by tapping the sparkle icon labeled "Make me move" from the top-right corner of the YouTube app's Create screen.
There are constraints on the input photo. It must show a single person. Group photos are not supported. The body needs to be fully visible, because the AI maps the reference movement onto the subject's pose. If the body is cropped at the waist or partially hidden, the output will either fail or produce visible artifacts.
The underlying model captures motion data from the reference video and transfers it onto the still image. Think of it as motion retargeting applied to a photograph rather than a 3D rig. The AI infers the subject's body structure from the photo, then warps and redraws frames to simulate movement that matches the reference clip.
The output is a short video clip, not a full-length Short. YouTube warns that results may be "inaccurate" because the tools are still experimental, which in practice means you should expect occasional limb distortion, clothing artifacts, and faces that drift slightly off-model between frames.
For teams that already repurpose still images across platforms, Make Me Move introduces a way to generate motion content without a camera or editing software. A product photo, a team headshot, or a behind-the-scenes still could become a Short with a few taps. But the output is constrained to the preset movements YouTube provides, so creative control is limited.
Can I upload my own reference video instead of using a preset? No. Make Me Move only supports the preset movement templates YouTube provides. You cannot supply a custom reference clip.
Does it work with photos of multiple people? No. The feature requires a single-person photo. Group shots are not supported and will either be rejected or produce poor results.
Why can I not see the feature in my YouTube app? If you are in the EEA, UK, or Switzerland, the feature is not available. It also requires an up-to-date version of the YouTube mobile app. Desktop and YouTube Studio do not support it.
How long is the output video? The output is a short animated clip. YouTube's image-to-video tools in Shorts generally produce clips of around six seconds, though exact duration may vary by preset.
Are the generated videos labeled as AI content? Yes. All Make Me Move outputs carry a SynthID watermark and an AI-generated content label, following YouTube's transparency policies for generative AI features.

Be the first to know when we're adding new features and releasing new updates!