You've probably seen a very realistic AI video saturated social media.
A stormtrooper who makes a snowman? Google Veo3. Is SurfingUnicorn Passing Ice flo made while penguins are rave reviews under the Aurora? Also, AI…we assume.
If you can dream, you can create it. This is extremely exciting, but it's incredibly unsettling in terms of misinformation and fact-checking about what it means for the creative industry.
Quick Rundown: With Google, you can create video clips of movies simply by entering what you want to watch. It contains realistic voices and sounds, which sets it apart from the rest of the models.
Given that I never had the skills to be a filmmaker, I was surprised that writing a few sentences on a computer would allow me to make clips of what previously required Hollywood special effects teams.
The video below shows three videos created with Metro to test the new technology Google launched in the UK on May 30th.
1. Office fantasy
First, I thought seeing what Londoners really think about the Metro would boost their team's morale.
I used this prompt.
Start with wide shots. A glorious sunny day. A quiet road. The truck to the excited crowd gathered at London's High Street Kensington is home to a majestic building that includes Whole Foods. The crowd is young, cool and full of anticipation. A red London bus pulls in the filming and the door opens. A brown Lakeland Putterdale Terrier has come off the bus and barked excitedly, jumping into the arms of a man nearby, tall, tanned curly hair and short tidy beard, extremely handsome, wearing a navy three-piece suit. A statuesque woman with a chignon comes out of the bus, stacks of metropolitan newspapers getting caught in her arms. She distributes them into the crowd that became ecs and soon begins to read with great enthusiasm. “Long Live Metro,” they all cry at once.
At first glance, it's pretty realistic (no?)
However, you don't have to be Sherlock Holmes to notice that the bus is not normally open through the front screen. Shop sign reads “The Whole Hoove.”
Also, the crowd we imagined represents a multicultural and cool London. This may be relevant as it is well known that AI can include biases in the data being trained.
After trying to improve the prompt by making it even more detailed, we still couldn't get a diverse crowd, but we got a “long live metro”, pronounced “live” rhyming with “diving”.
2. Vandalism explodes outside of Primark
I wanted to test how easy it is to incite tension by looking realistic and create something that can be spread as misinformation. So I asked to make the video look like it was filmed on the phone.
It should look like it was being photographed with a phone camera. The scene is typical of British high streets, with shops such as boots, Primark and Tesco. A young man wearing a balaclava runs into the view with a hammer and begins to destroy every store window and scream, “You're going to pay for this.” The woman with a shopping bag tries to stop him, but he pushes her away.
Given the recent riots that have affected the UK town, I wanted to go viral on social media and create something that could incite some angry responses about law and order.
I don't think anyone will be fooled by this opportunity.
The video returned without sound (issues that affected quite a few videos, come later), and the assailant Balaclava disappeared from the mid-hammer swipe on his face.
Shot with a glossy, high-resolution feel, it definitely didn't look like grainy user-generated content.
3. Enter Helmas
After finally starting to burn out about Turkmenistan's fiery “Gate of Hell” crater, this came to mind as a potentially cinematic background.
Night falls near the “Gate of Hell” Darbaza Crater in Turkmenistan. The light from the fire inside makes the dark sky shine. The woman in her 40s wears a protective suit, but she peers into the depths with her hair down, watching the flames flicker. She said: “I say this pit will burn out quickly. Before that happens, I'll take the fire home.” She then wraps around the edges.
This was my favorite and not too detailed, but it was the most successful prompt. I didn't see any obvious AI flaws right away (probably because it was just one person and wasn't that complicated to create). I think special effects can even belong to blockbuster movies.
I don't think I should be too pleased with myself as I literally did anything that required talent to create it.
However, it opens up a new pathway to explore anything you can imagine, so you're not surprised that the feature has gone viral.
Does Google see where technology is heading?
Given that AI is already accelerating at an uneasy rate, I asked Google where the tech headlines are in the future (ock laughs because I feel that I'm already hopelessly outdated because I can't count my fingers).
Matthieu Lorrain, creative lead at Google Deepmind, said subway: “We've already seen the VEO 3, which is used in everything from creating social quick clips to turning internal jokes into moving memes or quickly visualizing cool concepts. These are some of the major use cases we've seen since the feature launched on Gemini.
Here are some of the clips I created to showcase the features:
For now, one of the annoying parts of making videos is that they can't be edited. For example, you can't improve your clip and ask an animal-loving editor to ask you not to drop your dog. It's just that I can think of a new clip completely.
Lorain said: For now, it's a case of experimenting with language and trying to generate the video the way you want it to. This is trial and error, but it's also part of the fun thing! ”
Google is currently testing its ability to generate videos from images. This is one of the most in demand regarding the potential of AI video and is potentially of concern.
If you can upload images of real people, you can create a compelling deepfake with the possibility of spreading false information. But there are good reasons why you might want to do this.
Reddit founder Alexis Ohanian recently used another AI software, Midjourney, to share a tweet from a video generated from a photo of his mother hugging him. Explaining that he lost his mother 20 years ago and his family couldn't afford a video camera, he created a short animation to better imagine either of the shots had happened, so he didn't have any moving images to remember her.
People may also want to imagine themselves in situations like James Bond, or even more surprised, for more refined content about their society.
For now, even though this is technically possible, you cannot specify celebrities at written prompts and create videos using publicly available images (for legal and ethical reasons).
I asked Gemini for a video in which Kiel's Starge gave a speech outside Downing Street to warn him to watch the invasion of a shining radioactive hamster, but sadly blocked him from bringing this to Technicolor.
How do I create a video in Google Veo 3?
Currently only available to those with a subscription, which costs £18.99 per month.
Once accessed, you can either enter a prompt into Gemini, a rival to ChatGPT's company, or use Flow, designed for more serious AI film productions, allowing you to use consistent elements such as specific characters throughout the clip.
Users can create three clips per day to prevent the server from being overloaded.
To make a movie, you simply write a paragraph about what you want to show it, detailing the style and camera work, subject and script. Google provided a list of tips for successful prompts here.
What happened to the sound?
Google warns users about Flow that video may “not necessarily be healthy” as audio is still an experimental feature (so if this happens to you, it's not an issue with your speaker).
They said the speech could be better with slightly longer transcripts and could be muted for minors and causing subtitles.
“We're working on that,” they said.
Will you see this in the cinema soon?
Like all other industries, it's a safe bet that AI will shake up filmmaking.
Simply upload information about it and you can already generate realistic sound “podcasts” on any topic. You wouldn't be surprised if you could generate your own feature films quickly on your favorite topic without logging in to Disney Plus or Netflix at all. Certainly, quality is probably mixed together, and just uploading the latest bestselling manuscript can cause copyright issues.
Lorraine said: “As far as futures, we fully understand the possibilities of AI in filmmaking, as well as groundbreaking technology. We consider the emergence of these tools to be enablers, and we can help a new wave of filmmakers tell their stories more easily. By providing early access to the flows, we have a better understanding of how our technology can support and integrate their creative workflows.
The VEO 3 represents a huge advance in quality, featuring realism, 4K output, incredibly realistic physics and audio. Like other powerful creative tools, it's about rewarding practice. The more the prompt explains, the better the video will be. When it comes to making the most of Veo 3, think about encouraging people to learn to speak the Veo language. The more fluent and descriptive you express your vision, the better your video will be. ”
Please contact our news team by emailing webnews@metro.co.uk.
For more stories like this, Check the news page.
More details: Game Inbox: Are AI going to ruin video games?
Details: Front Mission 3: Remake updates graphics with AI Slops, fans aren't happy
More details: UK Watchdogs may force Google to change – what are they?
