OpenAI’s text-to-video model Sora is driving internet insane

In a groundbreaking announcement, Microsoft-backed OpenAI revealed its latest venture into the world of artificial intelligence with the introduction of Sora, a cutting-edge software designed to generate minute-long videos based on text prompts. The revelation on Thursday sparked a whirlwind of reactions across the internet, with users expressing both awe and amusement at the possibilities offered by this futuristic technology.

According to the company’s statement, Sora is currently available for red teaming, a process that involves identifying potential flaws in the AI system. Additionally, visual artists, designers, and filmmakers can use Sora to obtain feedback on the model’s capabilities, opening the door to a new realm of creative possibilities.

The statement highlighted Sora’s ability to create intricate scenes with multiple characters, specific types of motion, and precise details of both the subject and background. Notably, the software can generate multiple shots within a single video, providing a level of complexity that captivates the imagination of internet users.

As news of Sora’s capabilities spread, social media platforms were flooded with wild prompts, imagining scenarios ranging from the sublime to the absurd. “A giant cathedral is completely filled with cats. There are cats everywhere you look. A man enters the cathedral and bows before the giant cat king sitting on a throne,” one user challenged Sora and shared the resultant video on the microblogging platform X (formerly Twitter), the post amassing over 1.5 million views.

OpenAI’s CEO Sam Altman also took to X with a joyful announcement of the astonishing AI model. “We’d like to show you what Sora can do, please reply with captions for videos you’d like to see and we’ll start making some!” Altman offered graciously in a post that drew the most incredible responses.

From making bicycles race on an ocean “with different animals as athletes riding the bicycles” captured from a drone camera view to “a wizard wearing a pointed hat and a blue robe with white stars, casting a spell that shoots lightning from his hand and holding an old tome in his other hand,” Sora’s renderings have captivated all netizens.

However, the manner of requesting and lauding the text-to-video model’s renditions of reality and fantasy soon spawned a fresh wave of memes. Many X users poked fun at Sora’s knack for realising absurd prompts by sharing real-life bizarre video clips and jokingly attributing them to OpenAI’s latest invention.

One user penned, “Sora by OpenAI is just amazing. Here was my prompt: Create a video of a politician drinking water when faced with tough questions during an interview on national television,” and attached a clip of Indian Prime Minister Narendra Modi appearing at a talk show.

Many are milking this new meme format to share absurd clips from Indian films. “Sora by OpenAI is just amazing. Here was my prompt: create a video of a girl playing basketball with a football in tennis dress on a golf course,” humoured another X user, using a clip from director David Dhawan’s 1999 film Haseena Maan Jayegi, starring Sanjay Dutt and Govinda.

Forecasting the confusion that awaits the world as Sora and similar technologies take off, one X user’s tongue-in-cheek post of a ‘prompt’ and Indian cop film Singham’s fight sequence were mistaken for a genuine Sora response.

In the original post, the user described the prompt they supposedly fed the AI model as, “An agile Indian police officer, in denim jeans and a black polo shirt, steps out of a drifting car wearing aviators, fires at the tires of another car, causing it to flip over. Swiftly, he pulls out the driver just as the car flips over him.” Accompanying this description was a brief clip from Singham which prompted concern from Canadian psychologist and public personality Jordan Peterson. “Stunning. Goodbye Hollywood. It’s been nice knowing you,” Peterson admired the supposed creation of Sora.

Not everyone appears too impressed or humoured by this latest tech marvel. According to OpenAI, Sora is a work-in-progress, can confuse the spatial details of a prompt and have difficulty in following a specific camera trajectory. One exhibit of this ‘glitch’ can be seen in a viral video generated by Sora following “a cat waking up its sleeping owner demanding breakfast.”

In the video, the sleeping owner’s arm melds into the blanket while the cat grows a ‘fifth leg’ in an unnerving distortion. “Do y’all remember how f****** worked up people got because Elsa’s hair clipped through her shoulder in order to make this very good shot? Do you? And now people are coming in like ‘holy sh*t that 5 legged cat looks so real life!’” protested one critic.

Another combined humour with fear and posed, “Me in court watching as they play AI footage of a crime I never did.” “It’s entirely over. How is it going to be possible to ever trust what you see anymore? AI needs regulating desperately, and it needs to be banned in all commercial use,” echoed one detractor among many calling for a ban on or strict regulations of the emerging technology.

Have something to add to the story? Share it in the comments below.