Google has added a new feature to its Gemini generative AI model. You can now use Gemini to create short videos from your photos.
Gemini Veo 3
Google first launched its Gemini generative AI model back in December 2023. Then, in May this year, the company released Veo 3, a state-of-the-art video generation model. The videos produced by Veo 3 were very realistic. They also included synchronised audio, which was a first for any video generative AI model. Now, Google has utilized the power of Veo 3 to add photo-to-video capability to Gemini.
What can you do?
You can now use Veo 3 in Gemini to transform a photograph into an eight-second video, complete with a soundtrack. To use the feature, you simply select “videos” from the tool menu in the prompt box. You can then upload a photograph. Next, you can describe the scene and the audio you want the AI to add. The AI will then set to work creating your dynamic video, animating everyday objects to bring your still images to life.
Sample videos
Google has shared a compilation video on its YouTube channel, showing several different animations from the same base image. The image simply shows a cardboard box on a table in a room with plain white walls. You then see an edited montage of videos.
- The box opens and fires out confetti with a whoosh sound. Hands then reach in to take out some red shoes.
- The box opens, and a shiny metal elevator rises up before the doors open.
- The box opens and the camera tracks in to reveal a toy sailboat bobbing on very realistic water. You can see the inside of the box through the water.
- Hands open the box, and the camera tracks in to reveal a live music performance by a band complete with a cheering audience.
- In another video, the camera tracks over the open box to reveal a miniature cityscape with illuminated buildings, as well as streets filled with pedestrians and moving cars.
- One video shows the box being held up so you can see a hamster moving around a tiny kitchen.
- The last sample video shows the box filled to the brim with water. A man in swim trunks then leaps into the box, causing it to rock, splashing the water.
What we think
With this latest release, Google is establishing its Veo 3 video generative AI as the best that’s currently available. No doubt there will be lots of examples where the technology hasn’t worked as well as those shown in the sample video. However, whatever you think about generative AI, you can’t fail to be impressed by this latest showreel.
It’s also becoming more and more difficult to distinguish AI-generated video from actual shot footage. There are a lot of difficult conversations still to be had about the training and use of generative AI. The genie is well and truly out of the bottle.
Availability
The photo-to-video capability in Gemini is available now. However, you will need to have either a Google AI Ultra or Pro subscription to use it.
