How do I make use of Sora OpenAI’s newest tool for creating videos
The day before, OpenAI released its video generation model, Sora, to the general public. The announcement is made on the sixth day of OpenAI’s “Shipmas” event, which spans 12 days of technology releases and demonstrations. This is what you need to know and how to use the video model today.
What is Sora?
Sora is an extremely robust AI modeling of video that creates videos from text-based commands, creates animated images, or even remixes videos into new designs. OpenAI first showed the model in February; however, today marks the first time that the company has released it for wider usage.
What’s new with this version?
Sora’s primary goal is to create stunning videos using simple prompts identical to the model that was first shown in February. However, Sora OpenAI has worked to make the system more efficient and cost-effective prior to this release. There are a few brand-new options that stand out, and two of them are notable.
One of them is called the Storyboard. It lets you make multiple AI-generated videos and then arrange them on a timeline in much the same way you do with traditional video editors, such as Adobe Premiere Pro.
The other is a feed that functions as a kind of gallery for creativity. Users can submit their Sora-generated videos on the feed, look up the themes behind some videos, modify them, and more generally, OpenAI says.
What is the maximum you can accomplish with this?
You can create videos using text prompts, modify the appearance of videos, and alter elements using the tool Remix or assemble several videos using Storyboard. Sora offers preset styles that you can add to your videos. These include dark and moody film noir or cardboard and papercraft that provide a feel of stop-motion. You can also cut as well as loop the videos you create.
Who can benefit from it?
To make videos using Sora You’ll need to join one of OpenAI’s premium plans, which are either ChatGPT Plus ($20 per month) or ChatGPT Pro ($200 per month). Both subscriptions give access to additional OpenAI products in addition. Users who subscribe to ChatGPT Plus can produce videos that last up to five seconds in length and have a resolution as high as 720p. The plan allows you to create 50 videos a month.
Users who have a ChatGPT Pro subscription can create videos that are longer and higher resolution, with a maximum resolution of 1080 px and a time that is 20 seconds. Additionally, they can have Sora create five different versions of a video with a single command, which makes it possible to examine options faster. Pro users are restricted to up to 500 videos per month. However, they can also make infinite “relaxed” videos, which are not created at the moment but are scheduled for when traffic is low.
Both subscription levels enable you to create videos using 3 aspect ratios, namely horizontal, vertical, and square.
If you don’t subscribe and you don’t have a subscription, you’ll only be watching the feed of Sora-generated videos.
OpenAI has begun the global rollout of Sora today; however, it is taking longer to go live within “most of Europe,” the company announced.
Related: What is OpenAI?
What is it? Where do I find it?
OpenAI has shut down Sora out of ChatGPT. For access, visit Sora.com and log in using an account on ChatGPT Plus and Pro accounts. ( MIT Technology Review was not able to access the site as of press time, but a notice on the site stated that sign-ups were not being accepted due to “currently experiencing heavy traffic.”)
How’d we get here?
Many things have occurred since OpenAI first introduced Sora in February. Some other tech firms have launched tools for creating videos, including Meta Movie Gen and Google Veo. There’s also been a lot of criticism. For instance, artists who had earlier access to play around with Sora have leaked the tool in protest of the method in which OpenAI has been conditioned to use it for artists’ work with no right to compensation.
What’s next?
Like any new version of models, it is to see what measures OpenAI is taking to prevent Sora from being used for unlawful or illegal reasons, such as the production of fakes or deepfakes. In terms of safety and moderation, the OpenAI employee admitted that the company “might not get it perfect on day one.”
Another question to be considered is the amount of computing power or energy Sora will consume each time it makes the video. The process of creating a video takes more processing time and consequently energy than producing an average text response using the form of a program such as ChatGPT. The AI boom has been an energy drainer and is now an obstacle for tech companies trying to reduce their carbon footprint, and the widespread availability of Sora and other video-related models that are similar to it is likely to cause more harm.