(NewsNation) — Google’s latest AI tool can now bring photos to life, turning a single still image into an eight-second video with sound and cinematic flair.
The new feature, released Thursday, is available to Google AI Pro and Ultra subscribers in select countries, the company said in a statement.
To use it, go to gemini.google.com, upload a photo and then describe the scene as well as any audio instructions — like background music or dialogue. In just 1-2 minutes, Google’s AI assistant Gemini generates an eight-second, downloadable video with motion and sound.
Google dumping Assistant for AI-powered ‘Gemini’
“You can get creative by animating everyday objects, bringing your drawings and paintings to life or adding movement to nature scenes,” the company said.
In a demo video, Google shows how a still photo of a cardboard box can be transformed into a live concert scene complete with a cheering crowd.
Still image from an AI-generated video (right) and the original still photo (left). Credit: Google
The new photo-to-video feature is powered by Veo 3, Google’s video generation model which initially launched to a limited audience in May. Since then, over 40 million Veo 3 videos have been generated across the Gemini app and Flow, Google AI’s filmmaking tool.
Study: ChatGPT may be eroding critical thinking skills
As for safety concerns, Google said it’s taken “significant steps” to make sure video generation is “an appropriate experience.” For example, Gemini won’t generate content that includes excessive violence, threatens child safety, or spreads factual inacurracies that could cause real-world harm.
The company said all generated videos include a visible watermark to show they were AI-generated, as well as an invisible digital watermark.
Still image from an AI-generated video (right) and the original still photo (left). Credit: Google
That said, there are reportedly a few kinks that still need to be worked out.
When Bloomberg News tested the feature on the web version of Gemini, it wasn’t able to follow more complicated prompts, such as making a person in a photo breakdance. The output also changed the facial features, and sometimes even the race, of the photographed subject in multiple instances, the news outlet said.
The AI model is expected to improve in future updates.
