OpenAI unveils Sora: a revolutionary tool for generating video from text ✨

Cliquez ici pour lire en français

February 15, 2024 is sure to be a date to remember. It was on this day that OpenAI unveiled its new tool for generating video from text. And the results are impressive, to say the least.

Text-to-video generation 🖌️

Before getting to the heart of the matter, take a close look at the following videos. Some of them have been generated by OpenAI’s new tool, Sora, while others haven’t. Can you tell the difference? Can you tell the difference ? (Don’t cheat)

So, which videos do you think have been generated by artificial intelligence? The correct answer is: all of them. Yes, you read that right, all the videos you’ve just watched were generated by Sora, OpenAI’s new tool. Here are the prompts used to generate them:

• Video 1 : Reflections in the window of a train traveling through the Tokyo suburbs.
• Video 2 : A movie trailer featuring the adventures of the 30 year old space man wearing a red wool knitted motorcycle helmet, blue sky, salt desert, cinematic style, shot on 35mm film, vivid colors.
• Video 3 : A cat waking up its sleeping owner demanding breakfast. The owner tries to ignore the cat, but the cat tries new tactics and finally the owner pulls out a secret stash of treats from under the pillow to hold the cat off a little longer.
• Video 4 : A Chinese Lunar New Year celebration video with Chinese Dragon.

If you’d like to see even more Sora-generated videos, you can visit the OpenAI website by clicking here.

Sora: A master of visual creation ✨

Sora is capable of generating videos of unprecedented complexity. You can imagine scenes with a multitude of characters with precise movements, evolving in realistic, detailed settings. The model includes not only your instructions, but also the logic of the physical world to create striking interactions and animations.

With her keen understanding of language, Sora brings unique and endearing characters to life. Their facial expressions and interactions convey intense emotions and create a deep connection with the viewer. What’s more, Sora can handle multiple shots within a single video, while maintaining consistency of character and visual style.

A tool open to very few people 🔒

Sora is currently in the hands of OpenAI’s red team. It is examining critical aspects in terms of risks and potential impacts. At the same time, OpenAI has provided access to a number of visual artists, designers and filmmakers to gather their valuable opinions on the evolution of the model and its adaptation to the needs of creative professionals.

In addition, it is crucial for OpenAI to share its research progress at an early stage. This allows them to collaborate with experts outside OpenAI, incorporate their feedback and raise awareness of the emerging potential of AI.

What do you think of this new tool? And AI development in general? Discuss it in comments.

Source : OpenAI