With just one command, AI will create realistic videos like this

From the command and the Sora model just introduced by OpenAI, users will have a short video under one minute, with high fidelity.

After the " fever " called ChatGPT, OpenAI continued to stir up the global technology community when introducing the Sora model capable of creating short videos under one minute, with high fidelity with just a few lines of code.

On social network X, many users expressed their amazement at the quality of videos from the new AI model. Not only are the images realistic, many short videos also show a physical simulation close to reality.

'This could be the moment that makes people say 'wow' about AI ,' said Tom Warren, editor at The Verge.

Of course, if you look closely, you can still see the errors in the videos. The videos that OpenAI has released so far are all very short, under 30 seconds, and longer videos will have more errors. Still, with just a few short videos, many users have imagined the scenarios in which Sora can be applied, as the videos illustrate.

Besides creating videos from user descriptions, Sora can also create videos inspired by any photo or expand existing videos, or automatically complete missing frames.

Sora is the Japanese word for sky. The team behind the technology, including researchers Tim Brooks and Bill Peebles, chose the name because it ' evokes the idea of ​​limitless creative potential'.

The quality is too real

On its homepage, OpenAI says Sora can create videos up to 60 seconds long with highly detailed scenes, complex camera movements, and multiple characters with vivid emotions.

The Sora model, recently introduced by OpenAI, is capable of generating short videos under a minute, with high fidelity, with just a few lines of code. (Photo: OpenAI).

Below, the tech startup illustrates with a command that reads: 'A bustling snowy Tokyo. The camera moves through the bustling city street, following a few people enjoying the beautiful snowy weather and shopping at nearby stalls. Beautiful cherry blossoms are flying in the wind along with snowflakes.'

After processing, the AI ​​model returns an amazing video, with an unmistakable image of Tokyo city and a beautiful moment when snowflakes and cherry blossoms appear together in one frame.

Without looking closely, it would be hard to tell that this is an AI-generated video. The virtual camera, as if attached to a drone, follows a couple slowly walking through a street scene.

One of the passersby is wearing a mask. Cars rumble down the riverside road to the left and shoppers on the right wander in and out of a row of small shops.

Sam Altman, CEO of OpenAI, created a video for his followers asking for "Grandma's cooking lesson in her Tuscan-style kitchen." (Photo: Sam Altman).

The video from basic commands quickly attracted more than 30 million views on the X platform. In the comments section, many users expressed amazement at the authenticity of the video.

CNBC believes that video could be the next challenge for generative AI, after chatbots and image generators have successfully penetrated the consumer and business world.

Besides exciting AI enthusiasts, the new technology also raises serious concerns about the spread of fake news, as major political elections approach around the globe.

According to data from machine learning company Clarity, the number of AI-generated deepfakes increased by 900% compared to 2023.

Alarming memoir for the film industry

OpenAI, the company behind the ChatGPT chatbot and the Dall-E image-generating software, is just one of many big names in tech racing to perfect this instant video-generating model.

In February 2023, a company called Runway introduced a completely new AI text-to-video technology.

Visual effects artist Evan Halleck of the Oscar-winning film Everything Everywhere All At Once even admits that Runway's AI tools have optimized his work.

AI-generated video depicts 'a stylish woman walking down a street in Tokyo.' (Photo: OpenAI).

'I can cut out characters and place them neatly on a photo in minutes, compared to half a day ,' the expert shared.

Experts say the application of AI in the entertainment industry is growing. From the development of deepfake technology to AI being used to create scripts, artificial intelligence is gradually infiltrating film production.

'It's faster and cheaper than manual labor. In my opinion, VFX is a very time-consuming and labor-intensive process. So it's great that everything is automated,' he continued.

The New York Times suggests that AI could speed up the work of seasoned filmmakers while completely replacing less experienced digital artists.

Sora's footage is certainly impressive, but that's not all. Wired notes that the most surprising thing about the Sora model is the abilities it wasn't trained to do.

Specifically, Sora not only creates videos that meet the needs from user descriptions, but it also seems to have a clear grasp of cinematic language.

Runway's AI tools optimized the work of the visual effects team on Everything Everywhere All At Once. (Photo: A24).

Additionally, one feature in Sora that OpenAI's development team didn't reveal is the ability to create videos from a single image or a sequence of frames.

'This would be a really fun way to improve storytelling. You could draw exactly what you have in your head and then make it happen ,' said Tim Brooks, a research scientist on the project.

According to Bill Peebles, another researcher on the project, OpenAI is aware that this feature also has the potential to generate fake information and can be abused.