Google Launches New Ways to Transform Photos Into Videos Using AI
The main topic of discussion surrounding AI-powered video is, of course, deepfakes and disinformation . Google’s Veo 3 is capable of creating hyper-realistic videos that could fool many of us into believing they are real, which poses a serious danger. Are you sure that the video you just watched was actually recorded by someone? Or was it created entirely on a computer based on a simple request?
Not all AI-powered videos are so sophisticated, however. Some seem to focus on simple innovation, like Google’s photo-to-video converters. The company first released such a tool to paid subscribers back in April. Powered by Veo 2, the tool animated static images based on your input, turning them into short 720p videos up to eight seconds long. Unlike the viral videos created by the company’s newest model, Veo 3, these videos are more like a trick — an artificial “living photo,” so to speak.
Photo-to-video conversion feature coming to two new apps
Google is working to bring the feature to more apps, and thus more users. This month, the photo-to-video feature became available to paid users of Gemini , this time powered by Veo 3. (Despite the more advanced model, these videos are still short and at 720p.) Perhaps the biggest change is the ability to add sound to videos, which helps amplify the impact. Now, according to The Verge , the feature is coming to two Google products, both for free: Google Photos and YouTube Shorts.
These new features are again powered by Veo 2, so you’re getting the previous-generation video model here. (It looks like Veo 3 will be coming to YouTube Shorts this summer.) You can’t add your own prompts to the tool, either. Instead, you can choose from preset prompts or select the “I’m Feeling Lucky” option. That could mean turning a photo of a crosswalk light into a video of a dancing sign. ( That’s a real example from Google , by the way.)
Google Photos lets you create videos up to six seconds long, but YouTube Shorts lets you choose the length of your video. If you’re not happy with the results, you can try again or click the “Like” icon, which will help train the model. (There’s also a “Like” icon.)
Like all AI-generated results, Google’s photo-to-video tool warns that it’s experimental and that “results may be unexpected.” This is likely a front for Google, so if your animation creates something offensive or distasteful, it’s not the company’s fault. It’s AI: What Are You Going To Do? There should be no doubt about whether these videos were created by AI, as Google says each one will be generated with a visible label indicating its artificial origin. The videos will also contain an invisible SynthID watermark, which Google is implementing in all of its AI-generated products to help users distinguish real content from AI-generated content.
The new tools are currently available in Google Photos in the US for iOS and Android, and will be available in YouTube Shorts in the US, Canada, Australia, and New Zealand over the next week.
Who are these features intended for?
I admit I don’t use AI tools much (or at all), so it’s probably hard for me to judge the merits of such things. But I wonder: who is this for?
I think photo-to-video tools will be interesting for a couple more generations. Look! I animated the family dog. And now the drawing I made this morning moves! And now I’m bored.
There’s a novelty to these tools, but in my opinion, it quickly wears thin. While it’s fun to share these AI videos with friends every now and then, I’d much rather see a real photo than a fake animation created from one. When you try to justify this feature beyond novelty, it falls apart for me. Google gives the example of a photo of a smiling man: Veo 2 animates a man throwing leaves into the air. Not only does this animation have no connection to what’s happening in the original image, it’s also a false memory — something that never actually happened. Show me a static version of what actually happened .