Observe ZDNET: Add us as a preferred source on Google.
ZDNET’s key takeaways
- Google Omni goals to do for video what Nano Banana did for photos.
- Creators can construct movies from textual content, photos, audio, or video.
- AI avatars may assist creators, however increase belief issues.
Final week, Google introduced a brand new AI video functionality that can both assist creatives produce higher-quality movies extra simply, or vastly enhance the quantity of AI slop on YouTube. I am betting it’s going to be a mixture of each.
Google announced Gemini Omnia software that raises the flexibility to create video through AI to a completely new degree. The corporate in contrast this announcement to the extent of enchancment in AI picture technology achieved when it launched Nano Banana.
Additionally: Google I/O 2026: Everything announced
Nano Banana raised the bar significantly on what was doable with picture technology. Omni purports to do the identical with video. Omni has begun to roll out, however I have never had an opportunity to play with it.
Google described Omni as “the place Gemini’s capability to purpose meets the flexibility to create.” Apparently, in line with the corporate, “With Omni, you possibly can mix photos, audio, video and textual content as enter and generate high-quality movies grounded in Gemini’s real-world information.”
Though Omni is “beginning with video,” Google mentioned the brand new mannequin can “create something from any enter,” so presumably we’ll see different media sorts generated by the software inside due time.
Additionally: 6 Android Auto apps I wish I found sooner, because they make every drive easier
Omni can even be out there in mannequin tiers, beginning now with Gemini Omni Flash. The aptitude is coming to the Gemini app, Google Movement, and YouTube Shorts. It is not clear whether or not the net model of Gemini will assist Omni, or whether or not you will want to make use of the Movement interface through your browser.
There are some standout options that make this a really attention-grabbing providing.
Clone your self
I truthfully cannot resolve if that is going to be a standout characteristic, a really massive concern for privateness, or an untethered slop generator. The corporate mentioned you possibly can create movies “with your personal voice by utilizing Avatars, which create a digital model of your self so you possibly can generate movies that look and sound such as you.”
Additionally: I used Nano Banana 2 to make perfect sketchnotes: 5 lessons learned
As a daily producer of YouTube movies for my channel, I am intrigued. There have been instances once I needed to place out a video, however was having a nasty hair day, a nasty voice day, or a nasty angle day, and I simply did not need that to return throughout in video.
May I simply feed a script into my digital twin avatar and have RoboDave do the speaking? Would my viewers discover? Would they care? Would they hate it? Would I? Clearly that is an space worthy of experimentation, nevertheless it’s most likely not one thing I am going to use usually.
I do my YouTube channel, partially, to maintain my talking and presentation chops up. Foisting that work on a digital avatar would possibly scale back my workload, however it will additionally scale back my coaching and apply.
Google could be very cautious to say that it is incorporating its SynthID digital fingerprinting expertise in these movies, to allow them to be verified as having been produced with Omni. Google additionally mentioned, “Past the avatar characteristic, when it comes to enhancing movies to vary audio and speech, we’re nonetheless working to check this and higher perceive how we are able to carry this functionality to customers responsibly.”
Physics mannequin
A few of it’s possible you’ll bear in mind the early days of video video games, when characters behaved extra like ragdolls than objects within the bodily world. As video games received higher, they started to include physics fashions, so if one thing received shot, knocked again, or dropped, it did so in a matter in keeping with the physics of the item.
Omni now incorporates physics into the movies it creates. Google mentioned it has “an improved intuitive understanding of forces like gravity, kinetic vitality, and fluid dynamics.” It additionally makes use of Gemini’s information to “join language, imagery, and which means in ways in which go far past sample matching.”
Additionally: OpenAI’s new image watermarks make it easier to spot AI fakes – here’s how
The corporate mentioned Omni can construct detailed movies from quick prompts and might generate movies for issues like explainers that break down pretty complicated concepts. I do not doubt this. The evaluation capabilities of NotebookLM’s audio overview and video overview to have the ability to create explainers are astonishing. If a few of that expertise discovered its means into Omni, issues may get attention-grabbing rapidly.
I really fed advertising and marketing paperwork and spec sheets into NotebookLM and it produced explainer movies for varied options of my safety product that had been higher than something I may have accomplished by hand, particularly within the time it took. The visuals on the time weren’t nice, however having complicated options defined in a clear video in below half-hour was a force-multiplier for my product launch schedule.
Enter selection
One in every of Nano Banana’s early standout options was its capability to recontextualize a picture. For instance, I had it take an image of me strolling in a park and alter it so I used to be sporting one thing near an admiral’s uniform on the bridge of an aircraft carrier. Whereas it did not get the uniform fruit salad and brass fairly proper, it did handle to precisely reproduce my physique and face.
Additionally: I turned casual selfies into professional headshots with Gemini
Omni proposes to take that to video, turning picture, textual content, video, or audio right into a “cohesive output.” Proper now, the one audio it can settle for is voice recordings, however the firm mentioned it’s going to “roll out different kinds of audio inputs quickly.”
The corporate additionally mentioned you possibly can create scenes, match types, describe what you need in pure language, and get character consistency all through the video.
Conversational enhancing
One side of manufacturing movies I don’t take pleasure in is the enhancing course of. It is usually enormously tedious. However, with Omni, “Gemini Omni provides you a neater strategy to edit video – with pure language. Each instruction builds on the final. Your characters keep constant, the physics maintain up and the scene remembers what got here earlier than.”
Google additionally mentioned you possibly can change parts within the video. I can see an enormous profit if it is doable to import a video and have the editor take away obstructions or change objects and backgrounds. It is not clear how lengthy a clip could be, or precisely how a lot enhancing you are able to do with Omni on a given plan, however these potentialities are thrilling.
Additionally: Are Sora 2 and other AI video tools risky to use? Here’s what a legal scholar says
Two different transformations the corporate mentioned the brand new Omni can do are:
- Change particular issues, or change all the things. Your video turns into the place to begin for one thing you by no means may have filmed your self.
- Take a video you shot and simply ask Omni to vary what’s taking place. Edit the motion, add in new characters or objects, or rework a second into one thing sudden.
Moreover, Google hasn’t but specified video format or decision. Will this be an expert software that may deal with 16:9 movies in 4K or 8K decision, or is it meant to be a software for the YouTube Shorts technology?
When OpenAI launched Sora, it was a novelty. Whereas customers abused it (we gave Sam Altman blue hair and made him sing ZDNET’s praise), it by no means managed to be a software that helped an expert’s workflow.
Whereas producing AI avatar clones and changing objects is likely to be enjoyable, I am hoping this functionality is prolonged in order that it is usable both inside Remaining Reduce, Premiere Professional, and DaVinci Resolve, or a minimum of built-in sufficient that these instruments can use edits created by Omni.
It is doable. Omni’s options might be rolling out to enterprise prospects and builders through a Google API.
Additionally: OpenAI’s new image watermarks make it easier to spot AI fakes – here’s how
I am additionally curious if Omni will embed the little diamond watermark within the nook of its movies, prefer it does with Nano Banana’s generated photos. Whereas it is good to know a clip was generated by AI, such watermarking will get in the way in which of utilizing the AI as an expert software.
Will we see licensing tiers the place the watermark could be eliminated? Or will we see third-party instruments crop up that take away the watermark, whether or not Google needs you to or not? Time will inform.
Would you employ Google Omni to create a digital avatar of your self for movies you did not need to file in particular person? Tell us within the feedback beneath.
You may observe my day-to-day venture updates on social media. You should definitely subscribe to my weekly update newsletterand observe me on Twitter/X at @DavidGewirtzon Fb at Facebook.com/DavidGewirtzon Instagram at Instagram.com/DavidGewirtzon Bluesky at @DavidGewirtz.comand on YouTube at YouTube.com/DavidGewirtzTV.
