OpenAI Unveils Sora: Enthusiasm and Challenges

On February 15, OpenAI introduced Sora, a new generative AI model that transforms text into video. The tool has generated excitement on social media, though it requires significant refinement before a full launch.

Introducing Sora, our text-to-video model.

Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions. https://t.co/7j2JN27M3W

Prompt: “Beautiful, snowy… pic.twitter.com/ruTEWn87vf

— OpenAI (@OpenAI) February 15, 2024

Sora can generate videos up to 60 seconds long with resolutions up to 1080p based on simple text prompts. These videos can include multiple characters, specific types of motion, and precise details of objects and backgrounds.

The tool is based on research from GPT and DALL-E 3. It operates using a diffusion model—transforming an initial image into statistical noise and then gradually removing this noise.

Announcing Sora — our model which creates minute-long videos from a text prompt: https://t.co/SZ3OxPnxwz pic.twitter.com/0kzXTqK9bG

— Greg Brockman (@gdb) February 15, 2024

Developers acknowledge that Sora still has several shortcomings. It struggles to accurately model the physics of complex scenes, often confusing cause and effect.

“For example, a person might bite a cookie, but the bite mark might not appear,” OpenAI explains.

The tool also has issues with spatial details. The resulting video may not align with specified directions, mistaking right for left.

Currently, Sora is available to a “red team” of testers, as well as select designers, artists, and filmmakers.

Social Media Reaction

The tool has captivated social media, trending on X with over 173,000 posts.

To demonstrate the model’s capabilities, OpenAI CEO Sam Altman has begun accepting user requests for video generation. At the time of writing, he has shared a total of nine videos created by Sora.

https://t.co/uCuhUPv51N pic.twitter.com/nej4TIwgaP

— Sam Altman (@sama) February 15, 2024

AI experts have noted that Sora’s capabilities are “speechless.”

I don’t even know what to say…

These clips generated by OpenAI’s Sora model have me speechless.

We knew good AI text-to-video would come, but this quickly? Unreal.

We’re stepping into a new world.

Buckle up. pic.twitter.com/zP7b5fKw5x

— Mckay Wrigley (@mckaywrigley) February 15, 2024

According to Nvidia senior research scientist Jim Fan, Sora is much more than just another “creative toy” like DALL-E 3. He described it as a “data-driven physics engine,” as the AI model not only generates abstract video but also intuitively creates the physics of objects within the scene.

If you think OpenAI Sora is a creative toy like DALLE, … think again. Sora is a data-driven physics engine. It is a simulation of many worlds, real or fantastical. The simulator learns intricate rendering, “intuitive” physics, long-horizon reasoning, and semantic grounding, all… pic.twitter.com/pRuiXhUqYR

— Jim Fan (@DrJimFan) February 15, 2024

Alongside this, some users have expressed concerns that tools like Sora could exacerbate the deepfake problem.

OpenAI, the builders of ChatGPT, have built Sora. A text to video AI system that creates 60 sec videos. It’s not yet available but is causing significant concern around its deep fake potential use. The issue isn’t just video content but how it’s used. https://t.co/Qujz3TfeH2

— Ian McLintock (@ian_mcl) February 16, 2024

One user suggested that major social networks should consider built-in protection against realistic fakes, highlighting the threat of substituting video evidence of crimes.

Open AI launched #Sora that gives almost realistic text to video GAI:
1. Deep fake detection becomes super important against spread of misinformation (platforms like X should’ve in-built debunker of deep fakes)
2. Courts will have a hard time with video evidence as proof of crime

— TheShunyaLab (@theshunyalab) February 16, 2024

Another user called for De-AI reverse engineering technology to prevent misinterpretation and misuse of content.

This is amazing! But we badly need a De-AI reverse engineering technology to avoid misinterpretation and deepfake generation to avoid misuse.#Sora #OpenAI #AI #DeepFake https://t.co/hazxryNlHI

— Neeraj Chauriya?? (@neerajchauriya) February 16, 2024

Previously, OpenAI began testing a “memory” feature for the ChatGPT chatbot, which retains information discussed in conversations to enhance user interaction.

Meanwhile, the company is actively countering the illegal use of its products.