ChatGPT’s new AI image generation abilities continue to be one of the hottest topics in AI right now, as the feature helped OpenAI convert more users to the service. The AI image generation tool is so important that OpenAI is considering building a social network around it.
This is all part of OpenAI’s need to make ChatGPT more useful by adding new capabilities to the chatbot so users have everything they might need at their fingertips.
That trend is obvious across the industry. It’s not just ChatGPT that’s getting more features that are accessible with a single text prompt. Google is doing it too with Gemini, and the company just added an amazing new feature to its chatbot, assuming you have a Gemini Advanced subscription.
Gemini will allow Advanced subscribers to create mind-blowing AI videos using Google’s Veo 2 videos right from the chat box. Veo 2 support is also coming to Whisk, an AI image generation tool that Google unveiled in December.
Google explained in a blog post how everything will work. As expected, making stunning AI videos with Gemini Advanced is incredibly simple now that Veo 2 is integrated into the chatbot.
Veo 2 will be available as a model in the Gemini dropdown menu. Select it in Gemini, input your text prompt, and the AI will handle everything. Veo 2 will give you an 8-second clip at 720p (HD) resolution in 16:9 landscape format. You’ll be able to download the MP4 file and share it on social networks, including YouTube.
Google says there’s a monthly limit for videos without revealing what it is. However, Gemini Advanced users will be notified once they approach that limit. If Veo 2 goes viral as ChatGPT’s new image generation tool did, it’s likely that many Gemini Advanced users will soon find out what that limit is.
There’s no telling when Gemini users who are on the free plan will get the feature. For now, you’ll have to pay the $20/month Google One AI Premium subscription that gives you access to Gemini Advanced.
The feature will be available globally, on the web and mobile, which means you’ll be able to create AI videos with ease right from your phone.
Veo 2, seen in action in the examples Google offered (above and below), is described by the company as a “leap forward in video generation, designed to produce high-resolution, detailed videos with cinematic realism.”
“By better understanding real-world physics and human motion, [Veo 2] delivers fluid character movement, lifelike scenes, and finer visual details across diverse subjects and styles,” Google says.
There’s no question that some of the videos Google created to demonstrate Gemini’s new video generation feature look like genuine clips shot with an actual camera.
On that note, Google explains how it tackled safety for Gemini Advanced videos. First, the tool should not be able to generate content that violates Google’s policies. Hopefully, that means you can’t create misleading clips. Also, the 8-second limit should help with that.
Second, all Veo 2 images are marked with SynthID, a digital watermark placed in each frame. This indicates that the videos are made with AI if anyone’s looking. But it’s not good enough, as there’s no watermark present on the clips themselves to indicate their provenance.
Separately, Veo 2 is also coming to Whisk, the AI image generation tool that Google launched a few months ago. In Whisk, a Google Labs project, you can create AI images by giving the AI a subject, a scene, and a style. The AI makes an image for you.
Now that Veo 2 is coming to Whisk, you can instruct the AI via a text prompt to animate the image created with Whisk, as seen in the following clip. You get the same 8-second clips as with Gemini Advanced.
To access Veo 2 in Whisk, you’ll need a Gemini Advanced subscription. Head over to labs.google/whisk to get started.