Google has rolled out a new photo-to-video feature in Veo 3, its video generation model introduced in May 2025. The update allows users to convert static photos into eight-second video clips with audio using Gemini’s creative tools.
According to David Sharon, Multimodal Generation Lead at Gemini Apps, over 40 million videos have been generated through Veo 3 across the Gemini app and Flow in the past seven weeks. He noted that users have explored a wide range of creative themes — from modern takes on classic fairy tales to ASMR videos imagining the sound of slicing cooling lava.
How the Photo-to-Video Feature Works
To create a video from a photo, users need to:
- Select the ‘Videos’ option in the prompt box
- Upload an image
- Describe the scene and optionally include audio cues
Gemini then generates a short animated video from the input. Users can animate objects, add motion to natural landscapes, or bring hand-drawn images to life. The final output can be shared or downloaded.
Safety and Content Integrity
David Sharon emphasized that video generation within Gemini is designed with safety in mind. Google uses extensive “red teaming” techniques to test and address potential issues before release. The company also conducts ongoing evaluations to identify misuse risks and enforce content safety policies.
All AI-generated videos feature:
- A visible watermark
- An invisible SynthID digital watermark
Users are encouraged to use feedback tools, such as thumbs up/down buttons, to improve future updates and safeguard content quality.
Availability
The feature is gradually rolling out to Google AI Pro and Ultra subscribers in select countries. It is available on gemini.google.com and through Flow, Google’s AI video creation tool.