2024-02-16 02:06:00
The new artificial intelligence is not yet available to the general public. (X:OpenAI)
OpenAI, the renowned artificial intelligence research organization behind ChatGPT, has marked a new milestone in the evolution of the technology with the launch of Sora, a tool capable of generating videos from textual descriptions.
This advancement not only expands the horizons of digital content creation, but also offers fascinating insights into how the interplay between language and visuals can transform the way artificial intelligence is used today.
Sora from OpenAI is a new model that, although still in the testing phase, can already create videos of up to one minute. The company, led by Sam Altman, says Sora can make videos that look very real and follow exactly what users ask for, without losing quality in the images.
The AI is capable of producing videos with very specific prompts. (X:OpenAI)
Sora operates using a very similar principle to generators that convert text to image. This means that users can give very detailed instructions or “prompts” regarding what they want to see in the video.
These cues can be as specific as the user wishes, allowing them to include precise details regarding camera movement, scenery, people’s clothing, and other elements they wish to incorporate into the scene.
This offers great creative freedom, as the tool adapts to the given specifications to generate a video that corresponds as closely as possible to the user’s vision.
The tool includes very specific prompts. (X:OpenAI)
OpenAI shared several examples of videos and prompts, one of them was:
“The beautiful, snowy city of Tokyo is full of life. The camera moves along the busy city street, following several people enjoying the beautiful snowy weather and shopping at nearby stalls. Beautiful sakura petals fly with the wind along with snowflakes.”
Currently, Sora is being used exclusively by a select group of people who are part of OpenAI’s “red team”, with the main objective of identifying possible failures, damages or risks in the tool.
The “red team” is a specialized group whose main function is to evaluate and test systems, software or applications to identify vulnerabilities, flaws or potential risks. Their job is to think and act like potential adversaries, emulating the methods and tactics a real attacker might use.
The AI is still in the testing phase. (X:OpenAI)
Additionally, OpenAI has extended access to Sora to a small community of visual artists, designers and filmmakers to gather their feedback to improve and adapt the model, making it more useful for those in the creative field.
The company has decided to share early on the developments and advances of its research to open a feedback channel with people outside of OpenAI, thus allowing a better understanding of how the capabilities of artificial intelligence can be enhanced for benefit and advancement in various areas, including This time it offers the public a glimpse into what future AI technologies will be able to do.
OpenAI has noted that the videos they shared on the blog to announce Sora have not received any editing.
The image quality of the videos is promising. (OpenAI)
Taking this into account, it seems that the tool has great potential, although it is necessary to experiment with it personally to evaluate its true capabilities.
However, the company has also warned that, as with any other artificial intelligence tool, Sora has its own limitations and challenges.
For example, you might experience problems accurately representing physics in complex scenes or fail to identify specific cause-and-effect relationships between elements in a video.
However, we will have to wait a little longer to be able to access and test this new model directly.
1708066236
#Create #realistic #videos #text #Sora #choice #OpenAI #creator #ChatGPT
